I've built AI-integrated backend systems spanning RESTful services, agentic workflows, and RAG applications using FastAPI, PostgreSQL, LangChain, and LangGraph.
For a real rental business, I built a multi-agent operations system that automated CRM updates, invoice processing, payment follow-ups, document extraction, and webhook-driven workflows. I also added approval gates so exceptions reach human review rather than failing silently.
During my AI/ML internship at Era Foundation, I built end-to-end machine-learning pipelines on clinical datasets using Python, PyTorch, and scikit-learn. My work included preprocessing, model training, evaluation, responsible AI practices, model governance, and data integrity.
I've also trained a CNN to classify 14 lung diseases and built AudioBridge, a multimodal voice-first assistant for blind users using Ollama, Whisper, OCR, speech output, and retrieval-based context.
