Zeki is available for hire

Zeki Bilgin

Verified Expert in Engineering

AI and Computer Vision Engineer and Developer

Istanbul, Turkey

Toptal member since July 10, 2025

Bio

Zeki is a senior AI and computer vision engineer with a PhD in computer science from the City University of New York. He has extensive experience developing practical solutions for industrial projects. Zeki has authored over 30 scientific papers in machine learning, deep learning, computer vision, federated learning, privacy-preserving deep learning, cybersecurity, and telecommunications.

Portfolio

More Wow LLC

Computer Vision, Python, Optical Character Recognition (OCR), GitHub...

Univa Health

3D Pose Estimation, 3D Reconstruction, Machine Learning, Computer Vision...

Keyo

Siamese Neural Networks, Amazon Web Services (AWS), Deep Learning...

Experience

Artificial Intelligence (AI) - 13 years
Deep Learning - 8 years
Machine Learning - 8 years
Image Processing - 8 years
TensorFlow - 7 years
Computer Vision - 7 years
Federated Learning - 5 years
FastAPI - 2 years

Preferred Environment

Python, PyTorch, TensorFlow, Linux, Windows, Amazon Web Services (AWS), Large Language Models (LLMs), Retrieval-augmented Generation (RAG), LangChain

The most amazing...

...thing I've created is a novel method that uses machine learning to determine if code is vulnerable. It has received numerous citations.

Work Experience

Computer Vision & OCR Developer

2025 - 2026

More Wow LLC

Built a module that generates random building envelopes and forms room spaces based on given constraints such as the number of rooms and types, project area, etc.
Developed a computer vision pipeline to convert a raster floorplan to a vectorized plan.
Conducted research on open source AI-based computer vision segmentation models for floor plans, and adapted and integrated a feasible one for the project.

Technologies: Computer Vision, Python, Optical Character Recognition (OCR), GitHub, Artificial Intelligence (AI), Machine Learning, API Development, RAG Architecture, Data Extraction, Document Parsing, PDF, CAD, Workflow Automation, APIs, AI Pipeline, AIOps, Agentic AI, Software Architecture, Decision Modeling, AI Agents, API Integration, Document Processing, OpenAI, RAG Systems, Prompt Engineering, Artificial Intelligence as a Service (AIaaS), AI Chatbots, Chatbot Conversation Design, OpenAI SDK, AI Integration, Integration, Architecture, OpenAI GPT-4 API, Slack API, OpenAI API, Agentic RAG Systems, RESTFul APIs, ETL, CI/CD Pipelines, Full-stack, Claude API, Coding, AI Copilots, Databases, Classification, LangGraph, Decision Trees, AI Tools, Large Language Models (LLMs), Hugging Face Transformers, Synthetic Data Generation, Multimodal Models, Visual Language Models (VLMs), Hugging Face, AI Architecture, Cloud Architecture, Data Architecture, Data Pipelines, Solution Architecture, Stakeholder Management, Agentic AI Systems, Agentic Frameworks, LoRA training, ControlNet, ComfyUI, LoRa, Elasticsearch, Node.js, Rule-based Programming, Hyperparameter Tuning, Benchmarking, Vertex AI, Data Cleaning, Data Modeling, ETL Pipelines, Infrastructure, Linear Regression, Scaling, Statistical Analysis, Statistical Learning, Claude, Computer Vision Algorithms, Amazon Machine Learning, Gemini, CCTV, Video Surveillance, FaceNet, AI Agent Orchestration, LLM Integration, Cloud, Algorithms, AI Design, API Gateways, Microservices, REST APIs, Agentic Coding, Model Evaluation, Image Segmentation, Anthropic, Amazon Bedrock AgentCore, AI Programming, AI Automation, React, n8n, Pattern Recognition, Databricks, PySpark, You Only Look Once (YOLO), Fine-tuning, vLLM, AI Assistants, Workflow Automation & System Integration, LlamaIndex, AI Prompts, ChatGPT Prompts, Azure Databricks, Leadership, Amazon EKS, Data Preprocessing, Large Data Sets, Datasets, Quantitative Modeling, AI Modeling, IT Strategy, Web Applications, Model Deployment, LLM Fine-tuning, LLM Reasoning, Large Language Model Operations (LLMOps), Neural Networks, Scalable Vector Databases, Documentation, Multimodal GenAI, Open-source LLMs, Technical Architecture, Data Quality, JSON, Text to Image, Text to Image AI, Applied AI, Open Weights, Modeling, System Architecture, Data Scientist, Multi-agent Orchestration, Alembic, Automation, Data Visualization, Feature Engineering, Key Performance Indicators (KPIs), SciPy, Metrics, Dashboards, Scripting, Back-end, Back-end Development, AutoCAD, Feasibility Studies, Personally Identifiable Information (PII), Keras, Evaluation, PDF Scraping, Real Estate, Labeling, Geometric Modeling, Microsoft Copilot, Product Engineering, Git, Model Development, Monitoring, Attention to Detail, Code Review, Team Leadership, AI Engineering, Supabase, AI Model Integration, Data Taxonomy, Taxonomy, Cloud Platforms, Data, Data Governance, PowerPoint API, Supervised Fine-tuning (SFT), Tokenization, API Design, Performance, Performance Optimization, Testing, CSV, Cloud Infrastructure, Unstructured Data Analysis, Perplexity, AI Consulting, Auditing, Audits, Reporting, Reports, Training, Vibe Coding, AI Compliance, Compliance, Google Apps Script, Single Sign-on (SSO), Stakeholder Alignment, System Administration, Technology, Microsoft Copilot Studio, Startups, Legal Technology (Legaltech), Low Latency, Models, AI Evaluation

Senior Data Scientist

2024 - 2025

Univa Health

Developed weight biomarkers for non-invasive weight prediction from face images, which were used in healthcare software designed for eating disorder treatment.
Created machine learning and computer vision models for weight estimation from face images.
Developed visual model explanation methods such as saliency maps and Grad-CAM to improve the interpretability of deep learning models, specifically identifying which regions in images are most important for weight prediction.

Technologies: 3D Pose Estimation, 3D Reconstruction, Machine Learning, Computer Vision, MediaPipe, Amazon SageMaker, Predictive Analytics, Predictive Modeling, Time Series Forecasting, Generative Artificial Intelligence (GenAI), Image Generation, Facial Recognition, Facial Tracking, Stable Diffusion, GitHub, Multivariate Statistical Modeling, Statistical Modeling, Data Extraction, Software Architecture, Decision Modeling, Data Anonymization, Machine Learning (ML) APIs, Coding, Databases, Classification, Decision Trees, Clustering Algorithms, AI Tools, Hugging Face Transformers, Synthetic Data Generation, Multimodal Models, Visual Language Models (VLMs), Hugging Face, AI Architecture, Cloud Architecture, Data Architecture, Data Pipelines, Solution Architecture, Stakeholder Management, Bioinformatics, Clinical Research, Node.js, Rule-based Programming, Hyperparameter Tuning, Benchmarking, Data Cleaning, Data Labeling, Data Modeling, ETL Pipelines, Infrastructure, Linear Regression, Scaling, Statistical Analysis, Statistical Learning, Computer Vision Algorithms, Amazon Machine Learning, Claims, CCTV, Video Surveillance, FaceNet, LLM Integration, Cloud, Semantic Search, Algorithms, Model Evaluation, AI Programming, 3D, Volumetric Data, Open3D, Mesh Optimization, AI Automation, Graph Neural Networks (GNNs), Pattern Recognition, Eye Tracking, Healthcare, Fine-tuning, vLLM, Amazon EKS, Data Preprocessing, Large Data Sets, Datasets, AI Modeling, IT Strategy, Model Deployment, LLM Fine-tuning, LLM Reasoning, Large Language Model Operations (LLMOps), Neural Networks, Documentation, Open-source LLMs, Technical Architecture, Data Quality, JSON, Image to Vector, Applied AI, Open Weights, Video Analysis, Modeling, System Architecture, Data Scientist, Alembic, Automation, Data Visualization, Feature Engineering, Key Performance Indicators (KPIs), SciPy, Metrics, Dashboards, Scripting, Back-end, Back-end Development, Feasibility Studies, Personally Identifiable Information (PII), Keras, Evaluation, Clinical Trials, Labeling, 3D Geometric Analysis, Gesture Recognition, Geometric Modeling, Product Engineering, Git, Model Development, Monitoring, Attention to Detail, Code Review, Team Leadership, Large Language Models (LLMs), Artificial Intelligence (AI), AI Engineering, AI Model Integration, Data Taxonomy, Taxonomy, Cloud Platforms, Data, Data Governance, Unity Catalog, Supervised Fine-tuning (SFT), Tokenization, Embeddings from Language Models (ELMo), API Design, Performance, Performance Optimization, Testing, CSV, Cloud Infrastructure, Healthcare Software, Unstructured Data Analysis, AI Consulting, Auditing, Audits, Reporting, Reports, Training, Vibe Coding, AI Compliance, Compliance, Google Apps Script, Single Sign-on (SSO), CDISC, HL7 FHIR Standard, Stakeholder Alignment, System Administration, Technology, Startups, Models, AI Evaluation

Senior Computer Vision Engineer

2023 - 2024

Keyo

Trained, maintained, and evaluated Siamese neural networks to generate embeddings from palm images for biometric identification.
Designed, implemented, and deployed a liveness detection algorithm to embedded hardware.
Implemented a pose estimation algorithm to measure the roll, yaw, and tilt angles of hands from images.

Technologies: Siamese Neural Networks, Amazon Web Services (AWS), Deep Learning, Image Processing, Liveness Detection, Python, Linux, Computer Vision, Data Science, NVIDIA TensorRT, Artificial Intelligence (AI), Machine Learning Operations (MLOps), Video Processing, NVIDIA CUDA, Convolutional Neural Networks (CNNs), Image Analysis, SQL, AI Adoption, FastAPI, Amazon S3 (AWS S3), Image Classification, OpenCV, Residual Neural Networks (ResNets), 3D Pose Estimation, 6-DoF, Pose Estimation, Containerization, DevOps, Data Engineering, AI Model Training, Model Validation, Supervised Learning, Data Analytics, Data Analysis, Statistics, Package Distribution, Data Annotation, Scikit-image, Applied Statistics, Generative Artificial Intelligence (GenAI), Image Generation, GitHub, Data Extraction, PostgreSQL, Workflow Automation, APIs, Software Architecture, Recommendation Systems, Decision Modeling, API Integration, Vector Databases, Integration, Architecture, RESTFul APIs, ETL, CI/CD Pipelines, Data Anonymization, Coding, Databases, Classification, AI Tools, Hugging Face Transformers, Synthetic Data Generation, Hugging Face, AI Architecture, Cloud Architecture, Data Architecture, Data Pipelines, Solution Architecture, Stakeholder Management, Node.js, Rule-based Programming, Hyperparameter Tuning, Benchmarking, Data Cleaning, Data Labeling, Data Modeling, ETL Pipelines, Infrastructure, Linear Regression, Scaling, Statistical Analysis, Statistical Learning, Computer Vision Algorithms, Amazon Machine Learning, FaceNet, Edge Computing, Cloud, Semantic Search, Algorithms, AI Design, API Gateways, Microservices, REST APIs, Model Evaluation, AI Programming, AI Automation, Graph Neural Networks (GNNs), Pattern Recognition, Motion Capture, Fine-tuning, Leadership, Amazon EKS, Data Preprocessing, Large Data Sets, Datasets, AI Modeling, IT Strategy, Web Applications, Model Deployment, LLM Fine-tuning, LLM Reasoning, Large Language Model Operations (LLMOps), Neural Networks, Scalable Vector Databases, Documentation, Technical Architecture, Data Quality, JSON, Image to Vector, Applied AI, Open Weights, Video Analysis, Modeling, System Architecture, Data Scientist, Alembic, Embedded Systems, Automation, Data Visualization, Feature Engineering, Key Performance Indicators (KPIs), SciPy, Grafana, Metrics, Dashboards, Scripting, Identity & Access Management (IAM), Back-end, Back-end Development, Feasibility Studies, Personally Identifiable Information (PII), Keras, Evaluation, Labeling, Gesture Recognition, Geometric Modeling, 3D Cameras, Product Engineering, Git, Model Development, Monitoring, Attention to Detail, Code Review, Team Leadership, Large Language Models (LLMs), AI Engineering, AI Model Integration, Data Taxonomy, Taxonomy, Cloud Platforms, Data, Data Governance, Unity Catalog, Sequence Models, Supervised Fine-tuning (SFT), Tokenization, Embeddings from Language Models (ELMo), Fully Sharded Data Parallelism (FSDP), API Design, Performance, Performance Optimization, Testing, CSV, Cloud Infrastructure, Real-time Data Pipelines, Unstructured Data Analysis, AI Consulting, Auditing, Audits, Reporting, Reports, Training, Vibe Coding, AI Compliance, AI Security, Compliance, Google Apps Script, Single Sign-on (SSO), Stakeholder Alignment, System Administration, Technology, Startups, Low Latency, Models, AI Evaluation

Technical Leader | Senior AI/ML Developer

2020 - 2024

Arcelik

Developed YOLO-based object recognition models to identify items and food inside smart fridges with integrated cameras.
Created Edge AI solutions for home appliances, particularly quantized versions of object recognition models.
Built sentiment analysis models using NLP to rank customer feedback from one (not satisfied) to five (highly satisfied).
Contributed to shaping the company's AI strategy and roadmap.
Implemented autonomous path-following algorithms for a robot vacuum cleaner project in ROS, based on LiDAR and ToF sensor data.

Technologies: Python, Edge AI, Machine Learning, Computer Vision, Natural Language Processing (NLP), Data Science, Scikit-learn, XGBoost, NVIDIA TensorRT, Artificial Intelligence (AI), Machine Learning Operations (MLOps), Object Detection, Video Processing, NVIDIA CUDA, Convolutional Neural Networks (CNNs), Image Analysis, AI Adoption, Strategy, OpenCV, Residual Neural Networks (ResNets), Open Neural Network Exchange (ONNX), 3D Pose Estimation, 6-DoF, Pose Estimation, QR Codes, six degrees of freedom, DevOps, Optical Character Recognition (OCR), Data Engineering, Patent Applications, Patents, Technical Writing, Robotics, AI Model Training, Model Validation, Anomaly Detection, Quantization, Predictive Maintenance, Supervised Learning, Unsupervised Learning, Time Series Data, Sensor Data, Data Analytics, Data Analysis, Statistics, Package Distribution, Data Annotation, Scikit-image, Applied Statistics, Predictive Analytics, Predictive Modeling, Time Series Forecasting, Image Generation, Facial Tracking, GitHub, Multivariate Statistical Modeling, Statistical Modeling, Data Extraction, Software Architecture, AI Marketing, Chatbots, Technical Leadership, Platform Engineering, Integration, CI/CD Pipelines, Coding, Databases, Classification, Decision Trees, Clustering Algorithms, AI Tools, Synthetic Data Generation, AI Architecture, Cloud Architecture, Data Architecture, Data Pipelines, Enterprise Architecture, Solution Architecture, Stakeholder Management, Rule-based Programming, Hyperparameter Tuning, Benchmarking, MySQL, Data Cleaning, Data Labeling, Data Modeling, ETL Pipelines, Infrastructure, Linear Regression, R, Scaling, Statistical Analysis, Statistical Learning, Computer Vision Algorithms, Amazon Machine Learning, Scala, Edge Computing, Semantic Code, Cloud, Semantic Search, Algorithms, AI Design, AWS Lambda, Model Evaluation, Image Segmentation, C++, Embedded Linux, Qt, Amazon Bedrock AgentCore, Graph Neural Networks (GNNs), Pattern Recognition, You Only Look Once (YOLO), Motion Capture, Fine-tuning, Leadership, Data Preprocessing, Large Data Sets, Datasets, AI Modeling, IT Strategy, Model Deployment, LLM Fine-tuning, LLM Reasoning, Large Language Model Operations (LLMOps), Neural Networks, Documentation, Technical Architecture, Data Quality, JSON, Applied AI, Deep Reinforcement Learning, Reprise License Manager (RLM), HRL, Internet of Things (IoT), Energy, Industrial Control Systems (ICS), Industrials, Open Weights, Video Analysis, Modeling, System Architecture, Data Scientist, Embedded Systems, CAN Bus, Automation, Data Visualization, Feature Engineering, Key Performance Indicators (KPIs), SciPy, Dashboards, Back-end, Back-end Development, Feasibility Studies, Keras, Labeling, Gesture Recognition, 3D Cameras, Product Engineering, Git, Model Development, Monitoring, Code Review, Team Leadership, Team Management, Large Language Models (LLMs), AI Engineering, Data Taxonomy, Machine-learned Ranking (MLR), Cloud Platforms, Data, Data Governance, Unity Catalog, Sequence Models, Supervised Fine-tuning (SFT), Tokenization, Embeddings from Language Models (ELMo), Performance Optimization, Testing, CSV, Cloud Infrastructure, Real-time Data Pipelines, Unstructured Data Analysis, AI Consulting, Auditing, Audits, Mentorship, Reporting, Reports, Team Mentoring, Training, AI Compliance, AI Security, Compliance, Workshops, Workshop Facilitation, Stakeholder Alignment, System Administration, Technology, Collibra, Low Latency, Models, AI Evaluation

Experienced Security Researcher

2018 - 2020

Ericsson

Developed and implemented a novel method using machine learning to predict whether a code fragment contains certain vulnerabilities, which was published and has received numerous citations.
Created a privacy-preserving deep learning (LSTM) model for forecasting consumers' hourly electricity consumption, resulting in a published study that has received numerous citations.
Invented seven patents within one year that have been granted internationally in telecommunication, network security, software security, and machine learning.

Technologies: Machine Learning, Vulnerability Assessment, Network Security, Wireless Networking, Data Science, Forecasting, Demand Forecasting, Scikit-learn, Time Series Forecasting, XGBoost, Artificial Intelligence (AI), Convolutional Neural Networks (CNNs), AI Adoption, Strategy, Data Engineering, Patent Applications, Patents, Technical Writing, AI Model Training, Model Validation, Supervised Learning, Data Analytics, Data Analysis, Statistics, Applied Statistics, Time Series, Predictive Analytics, Predictive Modeling, Statistical Modeling, Data Extraction, Time Series Analysis, Reinforcement Learning, Multi-armed Bandits (MABs), Data Anonymization, Reinforcement Learning from Human Feedback (RLHF), Coding, Sales Forecasting, Classification, Prediction Markets, AI Tools, Synthetic Data Generation, AI Architecture, Data Architecture, Enterprise Architecture, Stakeholder Management, Rule-based Programming, Hyperparameter Tuning, Benchmarking, Data Cleaning, Data Labeling, Data Modeling, Linear Regression, Statistical Analysis, Statistical Learning, Scala, Edge Computing, Semantic Code, Semantic Search, Algorithms, AI Design, Model Evaluation, Audio Processing, Context Engineering, Pattern Recognition, Vehicle Routing, Vehicle Tracking Systems, Data Security, Data Preprocessing, Datasets, AI Modeling, IT Strategy, Neural Networks, Documentation, Technical Architecture, Applied AI, Deep Reinforcement Learning, Reprise License Manager (RLM), HRL, Internet of Things (IoT), Energy, Industrial Control Systems (ICS), Industrials, R, Modeling, System Architecture, Data Scientist, Data Visualization, Feature Engineering, Key Performance Indicators (KPIs), Keras, Labeling, Model Development, Code Review, Team Leadership, AI Engineering, ABAC, Data, Data Mesh, Role-based Access Control (RBAC), Data Governance, Unity Catalog, Sequence Models, Supervised Fine-tuning (SFT), Testing, CSV, AI Consulting, Mentorship, Reporting, Reports, Team Mentoring, Training, AI Compliance, AI Security, Compliance, ACORD Data Standards, Workshops, Workshop Facilitation, W3C, Stakeholder Alignment, System Administration, Technology, Collibra, Models, AI Evaluation

Postdoctoral Researcher

2017 - 2018

University of Nebraska - Lincoln

Developed mathematical models and conducted simulation experiments for a computational sociology project, analyzing information diffusion and social influence in social networks using machine learning techniques.
Implemented social network simulations and modeled interactions between actors and information diffusion.
Prepared a draft paper containing the findings, literature review, and other details of the project.

Technologies: Social Network Analysis, Simulations, AI Adoption, Data Engineering, Technical Writing, Applied Statistics, Coding, AI Architecture, Rule-based Programming, Statistical Analysis, Algorithms, Applied AI, Data Scientist, Feature Engineering, Model Development, Combinatorial Optimization, Data, Reporting, Reports, Technology, Models, AI Evaluation

Experience

Palm-based Biometric Identification

I created a solution that identifies people from their palms. Accurately identifying individuals from palm images with zero false positives and an acceptable level of false negatives was exciting. The solution was capable of tolerating variations in palm distance and orientation. It was also resistant to paper- and screen-based spoofing attempts.

Object Recognition for Smart Fridges

I developed both (i) YOLO-based object recognition models for smart fridges equipped with integrated cameras to be deployed in the cloud and (ii) quantized versions for deployment on edge AI devices. Additionally, I implemented several image processing steps, including adaptive smart cropping to eliminate irrelevant regions, image correction to remove lens distortion, color enhancement, and blur measurement for quality filtering.

Vulnerability Prediction from Source Code Using Machine Learning

https://ieeexplore.ieee.org/stamp/stamp.jsp?arnumber=9167194

I built a novel method for representing source code in a format suitable for machine learning models, and used this representation to build a vulnerability prediction solution. This work was published and has received significant attention.

Privacy-preserving Deep Learning for Load Forecasting on Smart Grids

https://ieeexplore.ieee.org/abstract/document/9024520

Implemented a practical privacy-preserving load forecasting solution using differential privacy techniques with the TensorFlow Privacy library. The results demonstrate that meaningful privacy guarantees can be achieved with varying levels of impact on forecasting accuracy. This solution presents privacy-utility trade-off values across different privacy levels.

Background replacement

I developed a semantic segmentation model for a jewelry business that replaces the background of diamond images. Since diamonds are transparent objects, it was a challenging task, but we achieved satisfactory results.

Digital Biomarker Development for Healthcare

I developed digital weight biomarkers to track eating disorder conditions. I developed computer vision models that non-invasively estimate weight changes from pairs of facial images captured during consecutive consultation sessions.

Literature Review on DeepFake Detection

A comprehensive review of state-of-the-art DeepFake detection solutions. I reviewed DeepFake detectors in the literature, including both video and voice detectors, and prepared a comprehensive report. These detectors are capable of identifying various types of video and audio manipulations, such as face swapping, lip syncing, and talking faces.

Sentiment Analysis

https://github.com/zekidsp/Text-classification

This project involved multi-class text classification as part of Natural Language Processing (NLP). I conducted exploratory data analysis on the Yelp Dataset, which contains 650,000 training samples and 50,000 test samples of customer reviews rated from 1 to 5. I performed data preprocessing, text vectorization, neural network design, training, and evaluation.

Autonomous Path Following

Path-following algorithm. I implemented autonomous path-following algorithms for a robot vacuum cleaner project at Beko. We used ROS and Gazebo for the development environment, and algorithms were developed based on LiDAR and ToF sensor data.

RAG Implementation

A RAG system with an ingestion pipeline that processes PDFs, slide decks, and lecture transcripts, splits them into chunks, and retrieves and re-ranks the most relevant content based on a given query. I designed and implemented the chunking strategy, as well as the ingestion and retrieval pipelines.

A Personalized Academic AI Coach

I developed a personalized AI coach using a knowledge graph and a multi-agent system. I implemented the AI coach as a FastAPI microservice and used Neo4j to build the knowledge graph.

The knowledge graph is populated from several external sources, including learning management system (LMS) platforms, and stores students' learning experiences, such as enrolled courses, concepts studied, exam grades, homework due dates, and homework grades.

The AI coach proactively interacts with students based on their individual learning data retrieved from the knowledge graph, enabling personalized guidance, recommendations, and interventions.

Education

2006 - 2010

PhD in Computer Science

City University of New York - New York, NY, USA

Certifications

JANUARY 2023 - PRESENT

SQL Server

Infotech Academy

Skills

Libraries/APIs

PyTorch, TensorFlow, Scikit-learn, XGBoost, OpenCV, API Development, Slack API, OpenAI API, Claude API, Hugging Face Transformers, Node.js, REST APIs, Open3D, React, PySpark, vLLM, SciPy, Keras, PowerPoint API

Tools

Open Neural Network Exchange (ONNX), Scikit-image, Amazon SageMaker, GitHub, CAD, GIS, Visual Language Models (VLMs), ComfyUI, Claude, ChatGPT, n8n, You Only Look Once (YOLO), Claude Code, Codex, AI Prompts, Microsoft Power BI, Amazon EKS, Reprise License Manager (RLM), Grafana, Azure OpenAI Service, Claude Agent SDK, Microsoft Copilot, Git, GraphRAG, Collibra, AutoCAD, MATLAB, Gazebo Simulator

Languages

Python, C++, SQL, R, Scala, TypeScript, Google Apps Script, SPARQL, OWL, RDF, Shapes Constraint Language (SHACL), JavaScript, C

Frameworks

MediaPipe, Multi-armed Bandits (MABs), LangGraph, Agentic Frameworks, Qt, LlamaIndex, Alembic

Paradigms

DevOps, Anomaly Detection, Model Context Protocol (MCP), ETL, Synthetic Data Generation, Rule-based Programming, Microservices, Automation, Machine-learned Ranking (MLR), ABAC, Role-based Access Control (RBAC), Testing, HL7 FHIR Standard, W3C, Siamese Neural Networks

Platforms

NVIDIA CUDA, Google Cloud Platform (GCP), Azure, Vertex AI, LangSmith, AWS Lambda, Embedded Linux, Azure Functions, Databricks, Kubernetes, Azure AI Search, Microsoft Copilot Studio, Docker, Linux, Windows, Amazon Web Services (AWS)

Storage

Amazon S3 (AWS S3), PostgreSQL, MongoDB, Databases, Data Pipelines, Elasticsearch, MySQL, JSON, Azure Cosmos DB, Azure Cloud Services

Industry Expertise

Applied Statistics, Bioinformatics, Healthcare

Other

Deep Learning, Image Processing, Computer Vision, Machine Learning, Natural Language Processing (NLP), Federated Learning, Artificial Intelligence (AI), Data Science, Forecasting, Demand Forecasting, Time Series Forecasting, Large Language Models (LLMs), NVIDIA TensorRT, Machine Learning Operations (MLOps), Object Detection, Video Processing, 3D Reconstruction, 3D Image Processing, Medical Imaging, Convolutional Neural Networks (CNNs), Image Analysis, AI Adoption, Strategy, Image Classification, Residual Neural Networks (ResNets), 3D Pose Estimation, 6-DoF, Pose Estimation, QR Codes, six degrees of freedom, Containerization, Optical Character Recognition (OCR), Data Engineering, Patent Applications, Patents, Technical Writing, Robotics, AI Model Training, Model Validation, Quantization, Predictive Maintenance, Supervised Learning, Unsupervised Learning, Time Series Data, Sensor Data, Data Analytics, Data Analysis, Statistics, Package Distribution, Data Annotation, Time Series, Predictive Analytics, Predictive Modeling, Generative Artificial Intelligence (GenAI), Image Generation, Facial Recognition, Facial Tracking, Stable Diffusion, Multivariate Statistical Modeling, Statistical Modeling, RAG Architecture, Data Extraction, Document Parsing, PDF, Workflow Automation, APIs, Retrieval-augmented Generation (RAG), LangChain, AI Pipeline, AIOps, Agentic AI, Software Architecture, Recommendation Systems, Decision Modeling, AI Agents, API Integration, Document Processing, OpenAI, RAG Systems, Prompt Engineering, Artificial Intelligence as a Service (AIaaS), AI Marketing, AI Chatbots, Vector Databases, Chatbot Conversation Design, Chatbots, Technical Leadership, Time Series Analysis, OpenAI SDK, Platform Engineering, AI Integration, Integration, Architecture, OpenAI GPT-4 API, Agentic RAG Systems, RESTFul APIs, CI/CD Pipelines, Reinforcement Learning, Full-stack, Data Anonymization, Multi-agent Systems, Reinforcement Learning from Human Feedback (RLHF), Vector Search, Pinecone, Machine Learning (ML) APIs, Coding, AI Copilots, Sales Forecasting, Classification, Decision Trees, RAG Pipelines, Clustering Algorithms, Prediction Markets, AI Tools, Multimodal Models, Hugging Face, AI Architecture, Cloud Architecture, Data Architecture, Enterprise Architecture, Solution Architecture, Stakeholder Management, Agentic AI Systems, Clinical Research, LoRA training, ControlNet, LoRa, Hyperparameter Tuning, Benchmarking, Data Cleaning, Data Labeling, Data Modeling, ETL Pipelines, Infrastructure, Linear Regression, Scaling, Statistical Analysis, Statistical Learning, Computer Vision Algorithms, Amazon Machine Learning, Claims, Gemini, CCTV, Video Surveillance, FaceNet, Edge Computing, AI Agent Orchestration, LLM Integration, Semantic Code, Cloud, Semantic Search, Algorithms, AI Design, Agentic Workflow Design, API Gateways, Distributed Systems, Agentic Coding, Automatic Speech Recognition (ASR), Model Evaluation, Image Segmentation, Audio Processing, Anthropic, Azure Function App, Amazon Bedrock AgentCore, AI Programming, Conversational AI, 3D, Volumetric Data, Mesh Optimization, Context Engineering, Trading, Fintech, AI Automation, Graph Neural Networks (GNNs), Pattern Recognition, Vehicle Routing, Vehicle Tracking Systems, Motion Capture, Eye Tracking, Cursor AI, Data Security, Meta Llama, Fine-tuning, Observability, AI Assistants, Workflow Automation & System Integration, AI Hallucinations Management, ChatGPT Prompts, Finance, Azure Databricks, Leadership, Speech-to-Text (STT), Data Preprocessing, Large Data Sets, Datasets, Quantitative Modeling, AI Modeling, IT Strategy, Web Applications, Model Deployment, Communication, LLM Fine-tuning, LLM Reasoning, Large Language Model Operations (LLMOps), Neural Networks, Scalable Vector Databases, Documentation, Multimodal GenAI, Open-source LLMs, Text-to-Speech (TTS), Technical Architecture, Data Quality, Text to Image, Text to Image AI, Image to Vector, Applied AI, Knowledge Graphs, Deep Reinforcement Learning, HRL, Internet of Things (IoT), Energy, Industrial Control Systems (ICS), Industrials, Open Weights, Video Analysis, Geometry, Conversational Agent, Modeling, System Architecture, Fraud Detection, Fraud Prevention, Data Scientist, Credit Scores, Multi-agent Orchestration, Embedded Systems, CAN Bus, Webhooks, Data Visualization, Feature Engineering, Key Performance Indicators (KPIs), Metrics, Dashboards, Scripting, Identity & Access Management (IAM), Microsoft Azure, Cloud Security, Back-end, Back-end Development, Feasibility Studies, Education Technology (Edtech), Personally Identifiable Information (PII), AI Voice Agents, Evaluation, HIPAA, SOC 2, PDF Scraping, Clinical Trials, Real Estate, Sentiment Analysis, Transcription, Labeling, 3D Geometric Analysis, Gesture Recognition, Geometric Modeling, 3D Cameras, Product Engineering, Model Development, Monitoring, Combinatorial Optimization, ChatGPT API, Attention to Detail, Code Review, Team Leadership, Team Management, AI Engineering, Answer Engine Optimization (AEO), Gemini API, Supabase, LLM Agents, OpenAI Agents SDK, AI Model Integration, Data Taxonomy, Taxonomy, Cloud Platforms, Data, Data Mesh, Data Governance, Unity Catalog, Sequence Models, Supervised Fine-tuning (SFT), Tokenization, Embeddings from Language Models (ELMo), Fully Sharded Data Parallelism (FSDP), API Design, Performance, Performance Optimization, CSV, Cloud Infrastructure, Healthcare Software, Real-time Data Pipelines, Unstructured Data Analysis, Perplexity, AI Consulting, Auditing, Audits, Mentorship, Reporting, Reports, Team Mentoring, Training, Data Analysis.Financial Data, Equity Trading, Investment & Trading Advisory.Stock Trading, Market Research.Market Research & Analysis, Vibe Coding, AI Compliance, AI Security, Compliance, Single Sign-on (SSO), Ontologies, ACORD Data Standards, CDISC, Semantic Web, Semantics, RDFs, Workshops, Workshop Facilitation, Stakeholder Alignment, System Administration, Technology, Startups, Legal Technology (Legaltech), Crypto, Contextual Bandits, Low Latency, Models, AI Evaluation, FastAPI, Research, Simulations, Wireless Networking, Liveness Detection, Edge AI, Vulnerability Assessment, Network Security, Secure Software Development Lifecycle (SSDLC), Differential Privacy, Semantic Segmentation, Speech Analytics, Social Network Analysis, Robot Operating System (ROS), LiDAR

Collaboration That Works

How to Work with Toptal

Toptal matches you directly with global industry experts from our network in hours—not weeks or months.

Share your needs

Discuss your requirements and refine your scope in a call with a Toptal domain expert.

Choose your talent

Get a short list of expertly matched talent within 24 hours to review, interview, and choose from.

Start your risk-free talent trial

Work with your chosen talent on a trial basis for up to two weeks. Pay only if you decide to hire them.

Top talent is in high demand.

Start hiring

Work Experience

Computer Vision & OCR Developer

2025 - 2026

More Wow LLC

Built a module that generates random building envelopes and forms room spaces based on given constraints such as the number of rooms and types, project area, etc.
Developed a computer vision pipeline to convert a raster floorplan to a vectorized plan.
Conducted research on open source AI-based computer vision segmentation models for floor plans, and adapted and integrated a feasible one for the project.

Senior Data Scientist

2024 - 2025

Univa Health

Developed weight biomarkers for non-invasive weight prediction from face images, which were used in healthcare software designed for eating disorder treatment.
Created machine learning and computer vision models for weight estimation from face images.
Developed visual model explanation methods such as saliency maps and Grad-CAM to improve the interpretability of deep learning models, specifically identifying which regions in images are most important for weight prediction.

Senior Computer Vision Engineer

2023 - 2024

Keyo

Trained, maintained, and evaluated Siamese neural networks to generate embeddings from palm images for biometric identification.
Designed, implemented, and deployed a liveness detection algorithm to embedded hardware.
Implemented a pose estimation algorithm to measure the roll, yaw, and tilt angles of hands from images.

Technical Leader | Senior AI/ML Developer

2020 - 2024

Arcelik

Developed YOLO-based object recognition models to identify items and food inside smart fridges with integrated cameras.
Created Edge AI solutions for home appliances, particularly quantized versions of object recognition models.
Built sentiment analysis models using NLP to rank customer feedback from one (not satisfied) to five (highly satisfied).
Contributed to shaping the company's AI strategy and roadmap.
Implemented autonomous path-following algorithms for a robot vacuum cleaner project in ROS, based on LiDAR and ToF sensor data.

Experienced Security Researcher

2018 - 2020

Ericsson

Developed and implemented a novel method using machine learning to predict whether a code fragment contains certain vulnerabilities, which was published and has received numerous citations.
Created a privacy-preserving deep learning (LSTM) model for forecasting consumers' hourly electricity consumption, resulting in a published study that has received numerous citations.
Invented seven patents within one year that have been granted internationally in telecommunication, network security, software security, and machine learning.

Postdoctoral Researcher

2017 - 2018

University of Nebraska - Lincoln

Developed mathematical models and conducted simulation experiments for a computational sociology project, analyzing information diffusion and social influence in social networks using machine learning techniques.
Implemented social network simulations and modeled interactions between actors and information diffusion.
Prepared a draft paper containing the findings, literature review, and other details of the project.

Experience

Palm-based Biometric Identification

Object Recognition for Smart Fridges

Vulnerability Prediction from Source Code Using Machine Learning

https://ieeexplore.ieee.org/stamp/stamp.jsp?arnumber=9167194

Privacy-preserving Deep Learning for Load Forecasting on Smart Grids

https://ieeexplore.ieee.org/abstract/document/9024520

Background replacement

Digital Biomarker Development for Healthcare

Literature Review on DeepFake Detection

Sentiment Analysis

https://github.com/zekidsp/Text-classification

Autonomous Path Following

RAG Implementation

A Personalized Academic AI Coach

Education

2006 - 2010

PhD in Computer Science

City University of New York - New York, NY, USA

Certifications

JANUARY 2023 - PRESENT

SQL Server

Infotech Academy

Skills

Libraries/APIs

Tools

Languages

Python, C++, SQL, R, Scala, TypeScript, Google Apps Script, SPARQL, OWL, RDF, Shapes Constraint Language (SHACL), JavaScript, C

Frameworks

MediaPipe, Multi-armed Bandits (MABs), LangGraph, Agentic Frameworks, Qt, LlamaIndex, Alembic

Paradigms

Platforms

Storage

Amazon S3 (AWS S3), PostgreSQL, MongoDB, Databases, Data Pipelines, Elasticsearch, MySQL, JSON, Azure Cosmos DB, Azure Cloud Services

Industry Expertise

Applied Statistics, Bioinformatics, Healthcare