Adil – Python, AI, Machine learning

At a leading sports technology platform, developed a real-time computer vision system to track players and equipment on the field, providing automated insights for performance analysis through deep learning models detecting movements, positions, and events.

• Developed sports analytics pipeline.
• Developed activity recognition model from scratch.

Architected an AI Legal Agent for a legal tech platform, leveraging LLMs, RAG, and intelligent document processing to ingest complex legal documents like contracts, NDAs, and regulatory filings—extracting key clauses, summarizing obligations, and providing accurate, cited answers to nuanced legal questions.

• Led design and end-to-end development of AI agent automating legal workflows.
• Built clause extraction, obligation summarization, and grounded Q&A with source citations.
• Implemented RAG pipeline including parsing, chunking, embedding for precise context retrieval.
• Created intelligent document processing to reduce manual review and boost decision confidence.

Led development of a computer vision system for a home-assistance robot that identifies and categorizes household clutter — toys, clothing, and objects on floors and furniture—in diverse domestic environments. Trained a DINOv2-based self-supervised model on 10M+ proprietary images using PyTorch. Architected an optimized training pipeline for an 8-GPU cluster, leveraging mixed-precision training and distributed data loading.

• Trained vision model on 10M+ annotated/unlabeled household images.
• Architected PyTorch training pipeline optimized for GPU cluster deployment.

Collaborated with a healthcare analytics company to clean and ingest complex electronic medical records from multiple provider networks using diverse systems and formats. Built a Python/SQL pipeline to parse, validate, normalize data — handling inconsistent schemas, missing values, duplicates, and PHI removal via regex/tokenization for HIPAA compliance. Delivered production-ready data warehouse powering clinical dashboards and predictive models, working closely with analysts and compliance teams for integrity and audit readiness.

• Analyzed sensitive client healthcare data from multiple sources.
• Ingested complex electronic medical records across diverse systems and formats.
• Standardized diagnosis codes, medications, encounter types, and ensured HIPAA-compliant PHI removal.