Aseem is currently unavailable
Aseem Gosain
Verified Expert in Engineering
AI/ML Engineer and Developer
Gurugram, Haryana, India
Toptal member since February 18, 2026
Bio
Aseem is a gold medal graduate of the Indian Institute of Technology and a senior data scientist with 5+ years of experience in applied AI and machine learning. His work spans classical methods, deep learning, and generative AI, with a particular focus on recommender systems, NLP, RAG, and LLM-based applications. Aseem drives projects end-to-end, from problem scoping and design through modeling, deployment, A/B testing, and stakeholder alignment, and collaborates effectively with global teams.
Portfolio
US-based Fintech Firm
Amazon Web Services (AWS), Claude Code, Docker, Python, PyTorch...
Simublade
Amazon Web Services (AWS), Python, PyTorch, Jupyter, Bash, ChromaDB...
Hyperqube Ionic
Python, Jupyter, Vector Databases, ChromaDB, NumPy, Pandas, Seaborn, Matplotlib...
Experience
- Python - 7 years
- SQL - 6 years
- Machine Learning - 5 years
- PyTorch - 5 years
- Amazon Web Services (AWS) - 5 years
- FastAPI - 3 years
- Pinecone - 3 years
- Claude Code - 2 years
Preferred Environment
Python, Amazon Web Services (AWS), PyTorch, FastAPI, Docker, OpenAI API, Vector Databases, SQL, Deep Learning, Claude Code, Artificial Intelligence (AI), Machine Learning
The most amazing...
...work includes building end-to-end product and merchant recommender pipelines for a major eCommerce marketplace, delivering measurable conversion rate lift.
Work Experience
Senior Data Scientist
2024 - PRESENT
US-based Fintech Firm
- Led the reliability, performance, and evolution of 12 production recommender pipelines, supporting product, brand, and merchant placement use cases across 12 million products, 8.2 million users, 7,000 merchants, and 50+ million interaction events.
- Designed, built, and maintained large-scale ETL and ML pipelines, using Dagster for orchestration, dbt for transformations, Amazon S3 for storage, and SageMaker for model training and inference, feeding a Redshift-based feature store and warehouse.
- Defined and implemented multi-strategy recommendation systems, including content-based, collaborative, two-tower retrieval, and GPT-driven recommender flows, and drove experimentation through A/B testing to improve personalization outcomes.
- Led the development of supervised ML classification solutions, building and deploying an XGBoost-based product category classifier to normalize marketplace data, reduce unclassified products by 80%, and align the catalog to a canonical taxonomy.
- Deployed ANN vector search encompassing Pinecone, ChromaDB, and Milvus, and set up Redash and Jupyter dashboards along with an A/B test harness, with early pilots showing 7% clickthrough rate (CTR) uplift on personalised offers versus control.
- Implemented agentic web scraping with OpenAI SDK, Playwright WebKit headless, and Dagster to automate product price audits across merchants, coordinating tasks with retries, tools, and memory to reduce manual effort by approximately 90%.
- Championed generative AI productivity tools such as Claude Code, Cursor, and ChatGPT, and best practices across a 10-person data science and ML team.
Technologies: Amazon Web Services (AWS), Claude Code, Docker, Python, PyTorch, Vector Databases, Recommendation Systems, Pinecone, Milvus, ChromaDB, Collaborative Filtering, Hugging Face Transformers, Amazon SageMaker, Amazon Redshift, Data Build Tool (dbt), Dagster, Redash, Cursor AI, Natural Language Processing (NLP), Model Context Protocol (MCP), OpenAI API, PostgreSQL, SQL, ETL, AI Pipeline, Data Pipelines, AI Agents, LoRa, Transfer Learning, CI/CD Pipelines, A/B Testing, Artificial Intelligence (AI), Machine Learning, Large Language Models (LLMs), Neural Networks, AI Programming, AI Tools, K-nearest Neighbors (KNN), Linear Regression, Regression, Ridge Regression, Lasso Regression, Model Regularization, Gradient Boosted Trees, Data Science, Business Intelligence (BI), Retrieval-augmented Generation (RAG), RAG Pipelines, Agentic RAG Systems, LangChain, LangGraph, AI Copilots, ChatGPT API, Claude, Claude Agent SDK, ETL Pipelines, Model Validation, Statistical Modeling, AI Assistants, Best Practices, Bots, Data Management, Claude API, CrewAI, AutoGen, Anthropic, Monitoring, Microservices, Generative Artificial Intelligence (GenAI), Software Development Lifecycle (SDLC), JSON, AI Model Training, Fine-tuning, Model Evaluation, Amazon Bedrock AgentCore, OpenAI, Prompt Engineering, Pydantic, Data Handling, Data Modeling, Python API, Data Engineering, Scientific Data Analysis, REST APIs, Cloud Services, Containers, Kubernetes, Data Analysis, APIs, Dashboards, Reporting, Analytics, Dashboard Design, Dashboard Development, Deep Learning, FAISS, API Development, Generative Pre-trained Transformers (GPT), RAG Systems, Technical Leadership
Senior AI/ML Engineer
2024 - 2026
Simublade
- Owned the architecture, development, and delivery of an AI-driven product matcher microservice using FastAPI and Docker for a confidential eCommerce retailer, integrating seamlessly with a .NET back end and Amazon EC2 infrastructure.
- Defined and implemented the core matching strategy, designing a hybrid ranking engine combining SBERT embeddings, Levenshtein distance, and regex-based rules to improve product matching accuracy by 50% while maintaining low-latency performance.
- Led production readiness and operationalization, establishing CI/CD pipelines with GitHub Actions, managing container registries, and implementing end-to-end monitoring dashboards to track throughput, latency, and failure rates across environments.
- Developed Python microservices using Flask and FastAPI, containerized with Docker for reproducible deployments across teams.
- Built production-ready ML and data pipelines and analytics warehouses using AWS S3, Redshift, SageMaker, Lambda, and EC2, as well as Snowflake SQL and dbt, optimizing for performance and reliability.
- Ensured scalability, reliability, and cross-team collaboration, supporting a distributed engineering team of approximately 10 developers with standardized deployment, observability, and release processes.
- Leveraged Copilot Studio, GitHub Copilot, Claude Code, and Cursor to write, optimize, and review code across multiple projects, improving developer productivity and code quality.
- Owned the end-to-end development of production ML/AI systems in Python, covering data ingestion pipelines, model training and inference, API services, and workflow automation, with a focus on clean architecture and production-grade performance.
Technologies: Amazon Web Services (AWS), Python, PyTorch, Jupyter, Bash, ChromaDB, Hugging Face Transformers, Transformers, TensorFlow, Cursor AI, SQL, Amazon SageMaker, Amazon S3 (AWS S3), Amazon EC2, AWS Lambda, GitHub, Recommendation Systems, Regex, Regular Expressions, Semantic Search, ETL, ML Pipelines, FastAPI, RESTful Microservices, Docker, Docker Compose, Artificial Intelligence (AI), Machine Learning, Large Language Models (LLMs), Neural Networks, AI Programming, AI Tools, K-nearest Neighbors (KNN), Linear Regression, Regression, Ridge Regression, Lasso Regression, Model Regularization, Gradient Boosted Trees, Data Science, Business Intelligence (BI), Retrieval-augmented Generation (RAG), RAG Pipelines, Agentic RAG Systems, LangChain, LangGraph, AI Copilots, ChatGPT API, Claude, Claude Agent SDK, ETL Pipelines, Model Validation, Statistical Modeling, AI Assistants, Best Practices, Bots, Data Management, Claude API, Anthropic, Monitoring, Startups, Microservices, Generative Artificial Intelligence (GenAI), Software Development Lifecycle (SDLC), JSON, AI Model Training, Fine-tuning, Model Evaluation, Amazon Bedrock AgentCore, OpenAI, Prompt Engineering, Pydantic, Data Handling, Data Modeling, Python API, Data Engineering, Scientific Data Analysis, REST APIs, Cloud Services, Containers, Data Analysis, APIs, Dashboards, Reporting, Analytics, Dashboard Design, Dashboard Development, Deep Learning, FAISS, API Development, Generative Pre-trained Transformers (GPT), RAG Systems, Technical Leadership
Lead ML Engineer
2024 - 2025
Hyperqube Ionic
- Owned the end-to-end design, development, and delivery of an AI-driven startup and investor matchmaking platform, including technical architecture, execution planning, and production readiness.
- Directed, mentored, and reviewed work for a team of four interns, setting coding standards, defining milestones, conducting regular reviews, and ensuring timely delivery of high-quality features.
- Defined and enforced clean, modular Python architectures for ML and AI applications, owning the full lifecycle from experimentation and rapid prototyping through production hardening, monitoring, and long-term maintainability.
- Designed and implemented large-scale data acquisition pipelines, building a multithreaded Selenium-based scraping system with undetected-chromedriver and humanized PyAutoGUI interactions to reliably collect data on 25,000 startups and investors.
- Designed hybrid retrieval pipelines using SBERT embeddings and vector databases (FAISS, Chroma, and Milvus) for efficient precision-latency trade-offs.
- Defined and built intelligent search and retrieval capabilities, developing a RAG-based knowledge bot using Jina and SBERT embeddings, ChromaDB, and Llama-70B hosted on Groq Cloud to support investor discovery and startup evaluation.
- Drove measurable efficiency and quality improvements, reducing manual research effort by around 70% while improving the relevance and accuracy of startup and investor matches through AI-assisted retrieval and reasoning.
Technologies: Python, Jupyter, Vector Databases, ChromaDB, NumPy, Pandas, Seaborn, Matplotlib, Recommendation Systems, Collaborative Filtering, Hugging Face Transformers, RAG Pipelines, RAG Architecture, Retrieval-augmented Generation (RAG), Agentic AI, Groq, Artificial Intelligence (AI), Machine Learning, Large Language Models (LLMs), Neural Networks, AI Programming, AI Tools, Linear Regression, Regression, Ridge Regression, Lasso Regression, Model Regularization, Gradient Boosted Trees, Data Science, Business Intelligence (BI), Agentic RAG Systems, AI Copilots, ChatGPT API, Claude, Claude Agent SDK, ETL Pipelines, Model Validation, Spatial Analysis, Statistical Modeling, AI Assistants, Best Practices, Bots, Data Management, Claude API, Anthropic, Monitoring, Microservices, Generative Artificial Intelligence (GenAI), Software Development Lifecycle (SDLC), JSON, AI Model Training, Fine-tuning, Model Evaluation, OpenAI, Prompt Engineering, Pydantic, Data Handling, Data Modeling, Python API, Data Engineering, Scientific Data Analysis, REST APIs, Cloud Services, Containers, Data Analysis, APIs, Dashboards, Reporting, Analytics, Dashboard Design, Dashboard Development, Deep Learning, FAISS, API Development, Generative Pre-trained Transformers (GPT), RAG Systems, Technical Leadership
IT Consultant
2021 - 2025
La Belle Sabbioneta Global SRL
- Led the design, development, and deployment of a Streamlit-based contactless ticketing app for a weekly church bus service, implementing QR-code-based validation and a lightweight SQLite back end to support seamless operations for 40 passengers.
- Defined and built predictive analytics solutions for the wine-tourism domain, developing models for traffic forecasting and dynamic price optimization to support data-driven business decisions.
- Delivered end-to-end analytics and visualization capabilities, creating Python-based dashboards that translated complex model outputs into clear, actionable insights for non-technical and business stakeholders.
- Acted as the primary technical owner, gathering requirements, selecting appropriate tools, and ensuring solutions were reliable, easy to operate, and aligned with real-world business workflows.
Technologies: Python, Streamlit, ChatGPT Prompts, ChatGPT, JupyterLab, Jupyter Notebook, Jupyter, SQLite, QR Codes, Web App Design, Forecasting, SARIMA, Time Series Analysis, Time Series Forecasting, AI Programming, Regression, Ridge Regression, Lasso Regression, Business Intelligence (BI), AI Copilots, Claude, IT Projects, Model Validation, Statistical Modeling, ARIMA, Best Practices, Bots, Data Management, Anthropic, Monitoring, Startups, Microservices, Software Development Lifecycle (SDLC), JSON, AI Model Training, Fine-tuning, Model Evaluation, OpenAI, Pydantic, Data Handling, Data Modeling, Python API, Data Engineering, Scientific Data Analysis, REST APIs, Cloud Services, Containers, Data Analysis, APIs, Dashboards, Reporting, Analytics, Dashboard Design, Dashboard Development, Deep Learning, API Development, Generative Pre-trained Transformers (GPT), Technical Leadership
ML Consultant
2022 - 2024
Univ.AI
- Executed end-to-end applied machine learning workflows for client-facing projects, including data cleansing, feature engineering, model selection, and training using TensorFlow and Keras, under academic guidance from Harvard faculty.
- Delivered production-grade ML and AI systems in Python, owning pipelines, models, APIs, and automation while optimizing performance, reliability, and scalability from prototype to enterprise deployment.
- Owned model optimization and adaptation efforts, fine-tuning pre-trained Hugging Face models through transfer learning on domain-specific datasets to improve task-specific performance and generalization.
- Collaborated in a consulting-style delivery model, translating client requirements into reproducible ML pipelines and validated model outputs within an Indian edtech apprenticeship environment.
- Recognized for exceptional technical aptitude and impact, earning a fully funded ML fellowship awarded to the top three candidates out of more than 5,000 applicants.
Technologies: Python, NumPy, Pandas, Scikit-learn, TensorFlow, Keras, PyTorch, Jupyter, Google Colaboratory (Colab), Matplotlib, Seaborn, Regex, SQL, SQLite, Pipelines, Convolutional Neural Networks (CNNs), Long Short-term Memory (LSTM), Recurrent Neural Networks (RNNs), Random Forests, Logistic Regression, Gradient Boosting, Decision Trees, Classification, Image Classification, BERT, Transfer Learning, Feature Engineering, Agile, Deployment, A/B Testing, Machine Learning Operations (MLOps), Generative Adversarial Networks (GANs), Reinforcement Learning, CSV, Plotly.js, Beautiful Soup, Requests, Artificial Intelligence (AI), Machine Learning, Large Language Models (LLMs), Neural Networks, Google Cloud Platform (GCP), AI Programming, AI Tools, K-nearest Neighbors (KNN), Linear Regression, Regression, Ridge Regression, Lasso Regression, Model Regularization, Gradient Boosted Trees, Data Science, Business Intelligence (BI), Retrieval-augmented Generation (RAG), AI Copilots, Claude, Claude Agent SDK, IT Projects, ETL Pipelines, Model Validation, Spatial Analysis, Statistical Modeling, AI Assistants, Best Practices, Bots, Data Management, Startups, Microservices, Generative Artificial Intelligence (GenAI), Software Development Lifecycle (SDLC), JSON, AI Model Training, Fine-tuning, Model Evaluation, OpenAI, Prompt Engineering, Pydantic, Data Handling, Data Modeling, Python API, Data Engineering, Scientific Data Analysis, REST APIs, Cloud Services, Containers, Data Analysis, APIs, Dashboards, Reporting, Analytics, Dashboard Design, Dashboard Development, Deep Learning, API Development, Generative Pre-trained Transformers (GPT), RAG Systems, Technical Leadership
Founder and Engineering Head
2018 - 2022
Vidyutm Mobilty
- Founded and led an electric vehicle hardware startup, owning the end-to-end development of a 20 kW three-phase motor controller and a 3.3 kW onboard battery charger for commercial van platforms.
- Directed core R&D and systems engineering efforts, defining hardware and software architecture, implementing CAN-based battery charging profiles using Raspberry Pi, and validating performance, safety, and reliability requirements.
- Owned supply chain, vendor management, and manufacturing readiness, sourcing components, coordinating prototyping, and balancing cost, quality, and lead-time constraints.
- Guided fundraising, regulatory, and policy engagements, working with government think tanks and ecosystem partners to align product design with evolving EV standards and incentive frameworks.
- Drove early commercial adoption and pilots, securing fleet-operator deployments and translating complex technical capabilities into clear, ROI-focused business value propositions.
Technologies: Python, Engineering, Electric Vehicles, Electrical Design, Control Systems, IT Projects, Statistical Modeling, Best Practices, Data Management, Startups, Software Development Lifecycle (SDLC), Data Handling, Data Modeling, Python API, Scientific Data Analysis, Data Analysis, Analytics, Dashboard Design, Dashboard Development, Deep Learning, Technical Leadership
Experience
Deep Learning-based Models for Malicious File Segregation in Naval Networks
https://doi.org/10.24868/11235I conducted research with my sister on using deep learning-based models for near-realtime malicious file segregation in naval networks with constrained resources and limited or no cloud connectivity. We proposed a hybrid model based on a recommender system to quickly segregate files using a CNN head and perform a deep scan of suspicious files using a transformer, CodeBERT. The research was presented and published in conference proceedings of Engine as a Weapon (EAAW) XI, 2025, held in Portsmouth, UK.
Issue Tracker RAG Bot
A template-filling assistant for construction site chat can convert chat conversations into structured, trackable tickets for the client. Based on BERT embeddings + Chromadb; prototype on AWS SageMaker studio.
Motion Detection Garden Cam
I built a prototype script to detect unsolicited guests—mainly cats and dogs—in my home garden using a webcam. It detects motion and forms bounding boxes around the source of motion. Additionally, it generates time-lapse footage of times of interest. The features it includes are OpenCV motion detection, a threaded video writer, timestamp overlay, and S3 archiving.
Education
2017 - 2018
Master's Degree in Power Electronics Engineering
Indian Institute of Technology (BHU) Varanasi - Varanasi, India
2013 - 2017
Bachelor's Degree in Electrical Engineering
Indian Institute of Technology (BHU) Varanasi - Varanasi, India
Certifications
AUGUST 2015 - PRESENT
Computer Science and Programming Using Python
edX
Skills
Libraries/APIs
PyTorch, NumPy, Pandas, Scikit-learn, Python API, REST APIs, TensorFlow, OpenAI API, Hugging Face Transformers, Matplotlib, Keras, Plotly.js, Beautiful Soup, Requests, OpenCV, PiLLoW, PIL, Claude API, Pydantic, API Development
Tools
Claude Code, Jupyter, ChatGPT, Claude, Seaborn, Amazon SageMaker, Redash, GitHub, SARIMA, LaTeX, You Only Look Once (YOLO), Claude Agent SDK, Docker Compose, ARIMA
Languages
Python, SQL, Regex, Bash
Paradigms
Best Practices, Microservices, Model Context Protocol (MCP), ETL, Agile, Business Intelligence (BI), Web App Design
Platforms
Jupyter Notebook, Amazon Web Services (AWS), Docker, Amazon EC2, AWS Lambda, Google Cloud Platform (GCP), CrewAI, Kubernetes
Storage
JSON, PostgreSQL, Amazon S3 (AWS S3), SQLite, Data Pipelines
Frameworks
Streamlit, LangGraph, AutoGen
Other
Machine Learning, Cursor AI, Pinecone, ChromaDB, Transformers, Natural Language Processing (NLP), Retrieval-augmented Generation (RAG), Vector Databases, Deep Learning, Recommendation Systems, JupyterLab, Google Colaboratory (Colab), Convolutional Neural Networks (CNNs), Long Short-term Memory (LSTM), Recurrent Neural Networks (RNNs), Random Forests, Logistic Regression, Gradient Boosting, Decision Trees, Classification, CSV, Regression, Linear Regression, Ridge Regression, Lasso Regression, Model Regularization, Gradient Boosted Trees, Artificial Intelligence (AI), Large Language Models (LLMs), Neural Networks, AI Programming, AI Tools, Data Science, AI Copilots, AI Assistants, Bots, Generative Artificial Intelligence (GenAI), AI Model Training, Fine-tuning, Model Evaluation, Prompt Engineering, Data Handling, Data Modeling, Cloud Services, Data Analysis, APIs, Dashboards, Analytics, RAG Systems, Engineering, FastAPI, FAISS, Milvus, Collaborative Filtering, Amazon Redshift, Data Build Tool (dbt), Dagster, ChatGPT Prompts, QR Codes, Forecasting, Time Series Analysis, Time Series Forecasting, Regular Expressions, Semantic Search, ML Pipelines, AI Pipeline, RAG Pipelines, RAG Architecture, Agentic AI, AI Agents, Groq, RESTful Microservices, LoRa, Transfer Learning, CI/CD Pipelines, A/B Testing, Pipelines, Image Classification, BERT, Feature Engineering, Deployment, Machine Learning Operations (MLOps), Generative Adversarial Networks (GANs), Open Source, Llama 3, Image Processing, Electric Vehicles, Electrical Design, Control Systems, K-nearest Neighbors (KNN), Agentic RAG Systems, ChatGPT API, IT Projects, ETL Pipelines, Model Validation, Statistical Modeling, Data Management, Anthropic, Monitoring, Startups, Software Development Lifecycle (SDLC), OpenAI, Data Engineering, Scientific Data Analysis, Containers, Reporting, Dashboard Design, Dashboard Development, Generative Pre-trained Transformers (GPT), Technical Leadership, Reinforcement Learning, LangChain, Spatial Analysis, Amazon Bedrock AgentCore
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.
1
Share your needs
Discuss your requirements and refine your scope in a call with a Toptal domain expert.
2
Choose your talent
Get a short list of expertly matched talent within 24 hours to review, interview, and choose from.
3
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