Nabeel is available for hire
Hire NabeelNabeel Raza
Verified Expert in Engineering
Machine Learning Engineer and Developer
Lahore, Punjab, Pakistan
Toptal member since October 4, 2021
Bio
Nabeel is a senior machine learning engineer with more than five years of expertise in AI/ML, deep learning, and natural language processing. He has a master's degree in data science, and his areas of expertise include computer vision, speech processing, and NLP. Nabeel has developed several projects to integrate ML models into applications. His programming language experience includes FastAPI, Flask, PyTorch, Keras, OpenCV, MongoDB, and many more libraries.
Portfolio
ProdKit.AI
Python 3, Python, Prompt Engineering, OpenAI GPT-4 API, LangChain...
Marketing Agency
Generative Pre-trained Transformers (GPT), OpenAI GPT-4 API, Python, LangChain...
Ramped
Python 3, FastAPI, MongoDB, Databases, Google Cloud Platform (GCP)...
Experience
- Python - 5 years
- Machine Learning - 5 years
- Data Science - 4 years
- Computer Vision - 4 years
- PyTorch - 4 years
- Deep Learning - 3 years
- OpenCV - 3 years
- Multiprocessing - 1 year
Availability
Part-time
Preferred Environment
Python 3, PyTorch, OpenCV, Scikit-learn, Pandas, Keras, Python, FastAPI, MongoDB, Cloud Firestore
The most amazing...
...thing I've done is publish a paper titled "k-Sparse Extreme Learning Machine" in a machine learning journal.
Work Experience
Founder
2024 - PRESENT
ProdKit.AI
- Developed a code explainer tool by utilizing OpenAI GPT and prompt engineering.
- Created APIs and handled data on the Firestore database.
- Created a chatbot for a project estimation report based on user inquiry.
Technologies: Python 3, Python, Prompt Engineering, OpenAI GPT-4 API, LangChain, Cloud Firestore, FastAPI, Large Language Models (LLMs)
GPT Wrapper Application Developer (via Toptal)
2024 - 2024
Marketing Agency
- Resolved issues within the existing tech stack and developed a proof of concept (POC).
- Utilized prompt engineering with LangChain to create objects, properties, records, and links.
- Integrated monitoring and the entire stack on AWS for comprehensive oversight and management.
Technologies: Generative Pre-trained Transformers (GPT), OpenAI GPT-4 API, Python, LangChain, AWS Lambda, AWS CloudFormation, API Gateways, Lambda Functions, ChatGPT Prompts, Prompt Engineering, Large Language Models (LLMs)
Senior ML Back-end Developer
2023 - 2024
Ramped
- Led a team of eight developers and created a job-searching platform from scratch.
- Developed a data pipeline to store and manage users' data in MongoDB and Firestore.
- Performed prompt engineering, integrated ML models in the platform, and created endpoints.
Technologies: Python 3, FastAPI, MongoDB, Databases, Google Cloud Platform (GCP), Generative Pre-trained Transformers (GPT), OpenAI GPT-4 API, Natural Language Processing (NLP), Large Language Models (LLMs), Stripe API, Firebase, Cloud Firestore, Python, Prompt Engineering, Hugging Face
ML Engineer
2022 - 2024
South Reno Athletic Club
- Developed a website that uses image processing and facial recognition using Python and Amazon Rekognition.
- Created FastAPIs to perform various website functions.
- Developed a dashboard to monitor the overall performance of the system.
Technologies: Computer Vision, Facial Recognition, Amazon Rekognition, APIs, Web Development, Algorithms, Deep Neural Networks (DNNs), Image Recognition, Neural Networks, Jupyter, Image Processing, Python 3, Python
Machine Learning Engineer
2023 - 2023
Global Tech Company
- Fine-tuned language models using DeepSpeed and Accelerate.
- Created a machine learning pipeline to train and validate LLMs.
- Performed text cleaning and other preprocessing techniques to improve performance.
Technologies: Language Models, Machine Learning, Fine-tuning, Causal Inference, DeepSpeed, Python 3, Python, Hugging Face, Large Language Models (LLMs)
Data Scientist | ML Engineer
2022 - 2022
Global Sales Automation Company
- Improved the accuracy by fine-tuning the existing human versus IVR classifier.
- Integrated Speech-to-text (STT) in the company's tech stack.
- Compared different speech models in terms of accuracy and time complexity.
Technologies: Python, Python 3, Keras, NVIDIA NeMo, Automatic Speech Recognition (ASR), Audio, Librosa, SciPy, NumPy, Pandas, Open Neural Network Exchange (ONNX), Speech to Text, Vertex AI, PyTorch, TensorFlow, Google Cloud Platform (GCP), Jupyter Notebook, Google Colaboratory (Colab), JupyterLab, Machine Learning, Deep Learning, Data Science, Speech Recognition, Artificial Intelligence (AI), Algorithms, Deep Neural Networks (DNNs), Vertex, Neural Networks, Time Series, Jupyter, Fine-tuning, Hugging Face
Machine Learning Engineer
2019 - 2021
Online Freelance Agency
- Ranked among the top AI and data science freelancers.
- Completed 20 projects in AI that involved image processing and training custom models, as well as fine-tuning models related to machine learning and deep learning.
- Worked on long-term contracts that included training and deploying efficient machine learning models.
Technologies: Python 3, PyTorch, OpenCV, Keras, Scikit-learn, TensorBoard, Pandas, Deep Learning, Data Science, Python, Machine Learning, Azure Machine Learning, NumPy, Matplotlib, TensorFlow, Azure ML Studio, Haystack, Computer Vision, Computer Vision Algorithms, Amazon Web Services (AWS), Multiprocessing, OSMnx, Generative Pre-trained Transformers (GPT), Natural Language Processing (NLP), Ubuntu, Flask, Voice Recognition, Artificial Intelligence (AI), APIs, Web Development, Object Detection, Object Tracking, Algorithms, Deep Neural Networks (DNNs), Image Recognition, Optical Character Recognition (OCR), Neural Networks, Time Series, Jupyter, Fine-tuning, Language Models, Image Processing, 3D
Research Associate
2018 - 2021
Lahore University of Management Sciences
- Published a paper in a machine learning journal titled "k-Sparse Extreme Learning Machine."
- Created a set of Python notebooks for an introductory course in machine learning.
- Fine-tuned a machine learning model for the diagnosis of osteoarthritis.
Technologies: Python 3, PyTorch, Keras, NumPy, TensorBoard, Python, Machine Learning, Azure Machine Learning, Scikit-learn, OpenCV, Pandas, Matplotlib, TensorFlow, Azure ML Studio, MATLAB, Data Science, Computer Vision, Computer Vision Algorithms, Deep Learning, Data Collection, Arduino IDE, Artificial Intelligence (AI), Algorithms, Deep Neural Networks (DNNs), Neural Networks, Time Series, Jupyter, Fine-tuning, Generative Artificial Intelligence (GenAI)
Data Scientist | ML Engineer
2019 - 2019
FiveRivers Technologies
- Performed data analysis and compiled reports on a ride-sharing platform dataset.
- Created a model that predicts price surges in the ride-sharing platform.
- Tested the company's custom automated machine learning software that creates multiple machine learning models on structured datasets.
Technologies: Python 3, Scikit-learn, Pandas, Python, Machine Learning, Azure Machine Learning, NumPy, Matplotlib, Data Science, Ubuntu, ETL, Deep Neural Networks (DNNs)
Intern
2016 - 2016
Packages
- Compiled reports about the working of packaging machines.
- Designed a machine that fetches plastic caps from a container.
- Used Arduino to move a robotic arm and fetch plastic caps.
Technologies: Arduino IDE
Experience
COVID-19 Sentiment Analysis on AWS
https://github.com/Nabeel965/covid_sentiment_analysisI created a complete machine learning pipeline in AWS for tweet sentiment analysis regarding COVID-19. Also, I employed AWS EC2 to scrape and store data in the AWS S3 bucket, preprocessed data using Amazon EMR, and created and validated models using Amazon SageMaker. Finally, the predictions were shown on a Tableau dashboard.
Dataset of Handwritten Words
https://www.kaggle.com/nabeel965/handwritten-words-datasetWe collected a dataset of handwritten English words from more than 100 people by asking them to write English words. This small dataset can be used for handwritten words recognition. It is publicly available on Kaggle.
Analyzing the Performance of Cricket Players
I developed a model that estimates a cricket player's performance in a future match based on weather, pitch type, and prior performance. In addition, I created a model that could predict the outcome of a game based on ground factors and the past statistics of both teams' players.
Voice Digits Recognition API
This API transcribes the audio containing digit voices. I created the neural network that predicts the five digits spoken in the audio and developed the Flask API that efficiently generates inference.
The Russian Language Speech-to-text Model
I trained the speech-to-text model using multi-GPU on a Russian telephone dataset, performed the telephonic augmentation, and created multiprocessing supported scripts for training and validation. I also used subtitles as the labels to train a model on Russian audio movies, which was made possible with the help of CTC.
Optical Character Recognition (OCR)
https://github.com/Nabeel965/OCR-PythonThis custom OCR, created using OpenCV in Python, converts images into English characters. I developed a script that takes an image input and predicts possible English characters in each line. I also trained connected component labeling (CCL) and convolutional neural network (CNN) to estimate characters.
Face Recognition Attendance System
A website I created enables instructors to mark the students' attendance by uploading an image of the classroom and the interference on low-quality images was improved using features like super-resolution (SR) imaging and image deblurring.
Kitchen Utensils Detection Using Raspberry Pi
https://github.com/Nabeel965/kitchen_utensils_detectionA model I created predicts the utensil type through the live camera mounted on a Raspberry Pi and I used OpenCV to provide fast interference on the live feed, where the accuracy of a lower FPS is improved by fine-tuning a deep learning model on a custom dataset.
Eye Segmentation and Color Detection
https://github.com/Nabeel965/Eye_color_detection/This Python script segments and predicts an eye color from a 3D image of a face. An eye is segmented using the connected component labeling (CCL) and the circular Hough Transform (CHT) algorithm in OpenCV.
Signature Extraction and Recognition
I created an API that takes in a PDF document and compares it to the signatures in the database. It is accomplished by extracting signatures from a PDF file using OpenCV and passing each candidate's signature through a machine learning model that matches it with the signature database.
Optimizing Pickup and Delivery Points in a Ride-sharing Platform
This API estimates the most likely pickup and delivery points in the future based on the history of pickup and delivery points. This estimation is accomplished by clustering previous data and calculating new points based on the time of day.
Stable Diffusion Fine-tuning and Optimization
I created a PyTorch-lightning version to fine-tune a stable diffusion model and the purpose of this project was to use lightning optimization techniques, where the model was fine-tuned on a custom dataset and on a machine with A100 GPU.
API Creation for Virtual Try-on Models
I compared different available try-on models and performed inference using PyTorch and Python with 3D image processing to extract the mask of the garment and I also created API using Flask and a demo on a web app using React.
Education
2018 - 2020
Master's Degree in Data Science
Information Technology University (ITU) - Lahore, Pakistan
2014 - 2018
Bachelor's Degree in Electrical Engineering
University of Engineering and Technology (UET) Lahore - Lahore, Pakistan
Certifications
FEBRUARY 2022 - PRESENT
AI Fellow
Pi-School
SEPTEMBER 2018 - PRESENT
Machine Learning Bootcamp
Evolve Machine Learners
DECEMBER 2017 - PRESENT
DeepLearning.AI Specialization
Coursera
MARCH 2017 - PRESENT
Machine Learning
Coursera
Skills
Libraries/APIs
PyTorch, Scikit-learn, NumPy, OpenCV, Pandas, Keras, Matplotlib, TensorFlow, Accelerometers, SciPy, Amazon Rekognition, PyTorch Lightning, React, DeepSpeed, Stripe API
Tools
Azure ML Studio, TensorBoard, Jupyter, MATLAB, Azure Machine Learning, Haystack, Open Neural Network Exchange (ONNX), ChatGPT, Praat, AWS CloudFormation, DeepSeek
Languages
Python, Python 3, Verilog, JavaScript
Platforms
Jupyter Notebook, Ubuntu, Amazon Web Services (AWS), Raspberry Pi, NVIDIA NeMo, Vertex AI, Google Cloud Platform (GCP), Firebase, AWS Lambda
Paradigms
ETL, Testing, Test-driven Development (TDD)
Frameworks
Flask, Accelerate
Storage
MongoDB, Databases, Cloud Firestore
Other
Machine Learning, Data Science, Computer Vision, Computer Vision Algorithms, Deep Learning, Speech Analytics, Neural Networks, Voice Recognition, Artificial Intelligence (AI), Algorithms, Deep Neural Networks (DNNs), Image Recognition, Optical Character Recognition (OCR), Fine-tuning, Image Processing, Generative Artificial Intelligence (GenAI), Time Series, APIs, Object Detection, Object Tracking, 3D, Data Collection, Multiprocessing, OSMnx, Arduino IDE, Natural Language Processing (NLP), Automatic Speech Recognition (ASR), Audio, Librosa, Speech to Text, Google Colaboratory (Colab), JupyterLab, Speech Recognition, Facial Recognition, Web Development, Vertex, Text to Image, Image Generation, Generative Adversarial Networks (GANs), Optimization, Hugging Face, Stable Diffusion, Language Models, Bittensor, ControlNet, Generative Pre-trained Transformers (GPT), Word Embedding, Causal Inference, Security, Ray.io, FastAPI, OpenAI GPT-4 API, Large Language Models (LLMs), DALL-E, Midjourney, LangChain, API Gateways, Lambda Functions, ChatGPT Prompts, Prompt Engineering
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