Languages
Python, SQL, JavaScript, TypeScript, Kotlin, Java, HTML/CSS
Toronto, Ontario, Canada
Om M. Patel
MSc CS (AI) at York University | NLP + Optimization
Vector AI Scholar at York University building NLP, optimization, and applied ML systems.
Open to AI/ML/NLP internships, research collaborations, and backend data engineering opportunities in Canada.
Skills
ML/NLP & LLM
PyTorch, TensorFlow, Scikit-learn, Hugging Face Transformers, LangChain, LlamaIndex, spaCy, FAISS, RAG pipelines
Optimization & Modeling
PyTorch Geometric, disjunctive-graph modeling, LoRA fine-tuning (4-bit), constrained prompting, GLM/GRU anomaly workflows
Robotics & Computer Vision
ROS2, OpenCV, YOLOv8, ArUco calibration, pixel-to-robot transforms, OCR with PyTesseract
Data Engineering & Backend
Airflow, Flask, REST API design, MongoDB, GeoPandas, Selenium, Beautiful Soup, ETL pipelines
Product, Deployment & Tooling
Docker, CI/CD, AWS EC2, Azure, Render, Next.js, Streamlit, Git/GitHub, Postman, PowerBI, Jira, Confluence
Education
Master of Science in Computer Science (AI Specialization)
York University, Ontario, Canada
Sep 2025 - Apr 2027
GPA: 3.63/4
Relevant coursework: Data Mining, Probabilistic Models and Machine Learning, Fairness and Bias in AI, Machine Learning Theory, Robotics
Bachelor of Technology in Computer Engineering
Pandit Deendayal Energy University, Gandhinagar, India
Nov 2021 - May 2025
GPA: 3.47/4
Relevant coursework: Computer Vision, Digital Image Processing, Pattern Recognition, Information Retrieval, Natural Language Processing, Big Data Analytics
Research & Publications
ADCIS 2024 (Springer): Automated Sleep Stage Classification Using Machine Intelligence Techniques
Co-author
Sep 2024 - Jul 2025
- Surveyed 90+ automated sleep staging studies and organized signal modalities and PSG data representations into an end-to-end taxonomy.
- Consolidated model families and evaluation practices with reporting guidance around data leakage, imbalance, and cross-subject generalization.
LLMs for Optimization: Job Shop Scheduling Optimization (JSSP)
Course Research Project, York University
Sep 2025 - Dec 2025
- Built a GNN-augmented LLM workflow by converting JSSP instances into disjunctive graphs and explicit machine/precedence constraints.
- Processed STARJOB-style data from ~130k raw records to 9,525 usable instances and reduced fine-tuning runtime from 70h to 11.5h with 4-bit LoRA.
Industry Experience
Backend & Data Engineering Intern
Uarra (Prev. Sharperly), Nigeria (Remote)
Dec 2024 - Aug 2025
- Designed Airflow-powered ETL pipelines for geospatial datasets, improving throughput by around 25%.
- Deployed Dockerized services on Render and Azure via CI/CD.
- Optimized MongoDB schemas and indexes for low-latency real-time lookups.
- Collaborated with ML and product teams to integrate production APIs and improve post-deployment reliability.
Data Science Intern
Uarra, Nigeria (Remote)
Mar 2024 - Jul 2024
- Built geocoding APIs in Flask + GeoPandas to reduce Google Maps API dependence and cut monthly cost by up to 45%.
- Scraped and processed 20k+ addresses to produce versioned datasets for ML training.
- Trained and evaluated RNN, Autoencoder, and K-Means models for user behavior prediction, route clustering, and anomaly detection.
Featured Projects
LLMs for Optimization Problems
Research-driven project exploring LLM + graph representations for constrained optimization (JSSP and related tasks).
OptimizationLLMJSSPGraph ML
GreenArm: Vision Guided Robotic Waste Sorting
ROS2 perception-to-manipulation pipeline using YOLOv8, OpenCV transforms, and ArUco calibration for precise robotic sorting.
RoboticsComputer VisionROS2YOLOv8OpenCV
PMML Project: Traffic Anomaly Detection in Toronto
Anomaly detection on urban traffic streams comparing probabilistic GLM and sequence-model (GRU) approaches.
Anomaly DetectionPMMLGLMGRUTime Series
Legal Clarity
Fine-tuned multilingual transformer summarization and translation pipeline for legal text, packaged as a Flask service.
NLPTransformersFlaskPyTorch
ChatPDF AI
Retrieval-grounded document QA system using LangChain and FAISS with agentic query workflows.
RAGLangChainFAISSAgentsLLM
Autograder: Precision in Every Grade
Semi-automated grading system with BERT-based NLP and OCR to evaluate typed and handwritten responses.
NLPBERTOCRReactFlask
Blog
Dec 18, 2025
LLMs for Optimization: What Worked in My JSSP Pipeline
My practical lessons from modeling Job Shop Scheduling with disjunctive graphs, constrained prompts, and LoRA fine-tuning.
OptimizationNLPMLOpsResearch Notes
Dec 5, 2025
PMML Project: GLM vs GRU for Traffic Anomaly Detection
My lessons from comparing Poisson GLM and GRU models, plus why KL-style divergence needs careful calibration in noisy traffic streams.
MLOpsResearch NotesOptimization
Nov 20, 2025
GreenArm Notes: Detection-to-Actuation Integration Pitfalls
My practical lessons from connecting YOLO detection, camera calibration, and Kinova arm control in a ROS2 waste-sorting pipeline.
MLOpsOptimizationResearch Notes
Oct 10, 2025
ChatPDF RAG Notes: Retrieval Quality Beats Model Size
How I tuned my ChatPDF pipeline by focusing on retrieval diagnostics, grounding checks, and failure slices before model tweaks.
RAGNLPMLOpsResearch Notes
Recruiter Fit Assistant
Paste a JD and get a quick fit brief for Om.
Certifications
Extracurriculars
Vector AI Scholarship Recipient, York University (May 2025)
Top-10 rank in Indian weather prediction at Kaggle ML Olympiad (Apr 2024 - May 2024)
AI/ML Mentor for 10+ students on projects including Autograder and Legal Clarity
Volunteer, Vardaan Foundation: taught digital literacy to underprivileged students (Jun 2022 - Jul 2022)