Hey there 👋 I'm Shahnoor, a Python-focused developer and data scientist with 5+ years of hands-on experience in agentic AI systems, LLM orchestration, web scraping, automation, full-stack development, and AI/ML systems. I build scalable, production-ready solutions that extract, process, and transform data into real business value. I've successfully delivered 40+ projects, scraped millions of records, handled captcha-protected and JavaScript-heavy websites, and built automation systems used in real-world workflows.

• Agent Orchestration
  • LangChain, LangGraph for multi-agent flows, stateful handoffs, and tool-calling chains
  • Custom orchestrators when frameworks add more complexity than they remove
• Local & Hosted LLMs
  • Ollama for self-hosted open-source models in data-sensitive environments
  • Frontier APIs (Anthropic Claude, OpenAI) with runtime routing logic based on input sensitivity and complexity
  • Finetuned open-source models for narrow, high-frequency tasks where general models are too slow or expensive
• Prompt Engineering
  • Structured output prompting, chain-of-thought with verification, dynamic few-shot selection based on retrieval
  • Constrained generation for tool-calling schemas, token optimization without losing reasoning quality
  • Versioned prompts tied to git commits, eval-driven iteration
• RAG & Vector Systems
  • Retrieval-augmented generation pipelines, pgvector, FAISS, ChromaDB, Qdrant
  • Grounded, auditable agent responses with observability at every step
• Infrastructure for AI
  • Async Python pipelines, parallel processing for multi-step agent flows
  • Dockerized LLM systems, CI/CD with eval suites as the test layer for non-deterministic behavior
  • MLflow, Weights & Biases for experiment tracking and model versioning

• Python & JavaScript
  • Expert in BeautifulSoup, Scrapy, Selenium, Playwright, Requests
  • Advanced handling of dynamic content, infinite scroll, pagination
• Anti-Bot & Scale
  • CAPTCHA bypassing, proxy rotation, session handling
  • Parallel processing, cron jobs, and large-scale crawlers
• Structured Data Delivery
  • Clean, normalized outputs in CSV, JSON, Excel
  • Ready for databases, analytics, or downstream pipelines

• LLMs & NLP
  • OpenAI API, LangChain, LangGraph, Hugging Face, SpaCy, NLTK
  • Attribute extraction, smart tagging, summarization at scale
• ML Frameworks
  • TensorFlow, PyTorch, Scikit-learn
  • Feature engineering, data cleaning, model training
• Automation with AI
  • AI-assisted data research
  • Intelligent workflows using LLMs

• Backend
  • Python: Django, Flask, FastAPI
  • Node.js, REST APIs, GraphQL
• Frontend
  • React, Next.js, HTML, CSS, Tailwind
  • Responsive, user-focused UI design
• Databases
  • MongoDB, PostgreSQL, MySQL, SQLite, Firebase, Redis

• Cloud Platforms
  • AWS, Google Cloud, Microsoft Azure
• Deployment & Infra
  • Docker, GitHub Actions, Render, Railway
• Workflow Automation
  • n8n, Make.com, Zapier, Webhooks
  • Windows Server scripting, cron-based pipelines

• ⭐ Top Rated on Upwork with 100% Job Success
• 🚀 40+ successfully delivered projects
• 📊 Scraped and processed millions of records
• 🔁 Built systems for parallel scraping & large-scale automation
• 🧠 Combined scraping + AI for smarter data extraction
• 🤖 Built production multi-agent pipelines using LangGraph with stateful handoffs and tool-calling
• 🔒 Deployed self-hosted LLMs via Ollama for data-sensitive workloads
• 📝 Versioned prompt engineering with eval-driven iteration and CI/CD for non-deterministic systems

• 💼 Upwork: Top Rated Python & Automation Developer [https://www.upwork.com/freelancers/shahnoorr]
• 🧑‍💻 GitHub: Exploring scalable scraping, automation & AI systems
• 🌍 Open to freelance, contract, and long-term collaboration

I enjoy solving hard problems, optimizing systems, and building tools that actually get used.

MIT