Owais I.Private profile: Upwork shows its earnings only to signed-in members, so we check it every few weeks, not daily. MRR here is a rough estimate (Upwork's six-month earnings ÷ months active). Last read Nov 5, 2026.

Need an AI agent grounded in your data, a RAG pipeline you can trust, or a dependable extraction workflow? I build Python systems that retrieve, reason, validate, and turn messy information into usable outputs. I am a PhD scholar working on legal AI and a university lecturer with 5+ years of teaching and applied AI/NLP research. I have completed 15 Upwork projects in model training, data extraction, forecasting, Python automation, and technical research. Clients have highlighted collaboration, reliability, technical knowledge, and on-time delivery. AGENTIC AI & RAG - LLM applications, multi-step agent workflows, tool use, and human approval gates - RAG pipelines with chunking, retrieval, citations, and measurable evaluation - Prompt versioning, bounded context, structured outputs, audit trails, and failure handling - FastAPI backends, React interfaces, Docker, tests, and GitHub Actions KNOWLEDGE GRAPHS & DOCUMENT AI - Neo4j knowledge graphs built from long, messy documents - Entity and relation extraction, graph-based retrieval, and citation-grounded answers - Structuring legal and policy text, including detecting conflicts between documents NLP & MACHINE LEARNING - Text classification, sentiment and offensive-language analysis, information retrieval, and generation - PyTorch, TensorFlow, scikit-learn, NLTK, spaCy, transformers, and word embeddings - Research design, dataset preparation, model evaluation, and technical documentation DATA EXTRACTION & ANALYSIS - Python, APIs, HTML/JSON/CSV/Excel processing, and document parsing - Cleaning, normalization, validation, and exploratory analysis - Repeatable workflows with clear outputs, error handling, and documentation SELECTED SYSTEMS - SupportOps AI: agentic customer-support workflow with policy retrieval, risk controls, structured LLM output, approval gates, and audit events - RAG Reliability Engine: BM25 retrieval, stable citations, and regression metrics for hit rate, MRR, precision@k, and citation coverage - ResolveAI Agent Arena: bounded decide-validate-act-observe loop with sandboxed tools and repeatable evaluation cases I scope the input, required output, risks, and acceptance criteria before building. You receive clear milestones, tested deliverables, documentation, and honest communication. Send me a sample input and the result you need. I will suggest the most practical first milestone.