I build and deploy complete applications: backend architecture, APIs, databases, auth, and polished frontends, with production AI features integrated end to end.
I'm a BS Software Engineering graduate who ships complete, deployed applications, not notebook demos. I work across the MERN/PERN stack with TypeScript and Python, and what sets my work apart is integrating AI features that are built to be reliable: RAG pipelines, self-correcting agentic tool-calling loops, and schema-validated LLM outputs.
- 🔭 Open to opportunities: Full-Stack (MERN / PERN) and AI-augmented engineering roles
- 💼 Experience: AI Engineering Intern at Developers Hub Corporation (Jul – Sep 2025)
- 🎓 Education: BS Software Engineering, Khawaja Fareed University of Engineering and Information Technology
- 🏅 Certifications & Community: Oracle Certified Foundations Associate · Microsoft Learn Student Ambassador (2023–2024)
- 💬 Ask me about: Agentic LLM tool-calling, multilingual RAG, PostgreSQL +
pgvector, TypeScript/Node.js backends, auth and role-based access control - 📫 Reach me at: [email protected]
|
A GPT-based function-calling agent runs a multi-round tool loop over four decision tools (spending limits, duplicate detection, receipt requirements, user history) to produce Zod-validated verdicts. Validation errors are fed back into the agent's prompt, so it corrects itself instead of failing silently. Includes receipt upload with OCR (Cloudinary + Tesseract.js), Clerk authentication, and role-based dashboards.
|
|
A full-stack RAG platform for legal contract analysis in English, Spanish, and Urdu: grounded Q&A with chunk-level citations, risk summaries, and clause extraction. Uses Gemini
|
AI Engineering Intern — Developers Hub Corporation (Jul 2025 – Sep 2025)
- Built and deployed a context-aware chatbot (LangChain, Google AI Studio, Chroma) that retrieves answers from a custom knowledge base and keeps context across multiple turns, shipped end to end via Streamlit.
- Built a production-ready churn prediction pipeline (Scikit-learn) covering preprocessing, training, and hyperparameter tuning; fine-tuned a BERT-base model for news topic classification, deployed via Streamlit/Gradio.
"Any fool can write code that a computer can understand. Good programmers write code that humans can understand."
— Martin Fowler
Let's build something reliable and scalable together. 🚀