Building agentic systems, scalable backend pipelines, and production-ready AI applications.
AI-first software engineering enthusiast focused on building agentic systems, scalable backend pipelines, and rapid prototypes. Skilled in LLM-based automation, GenAI application design, and full-stack development. Strong interest in high-impact engineering challenges, fast iteration cycles, and developing secure, production-ready AI systems.
Agentic RAG pipeline using LangGraph and Groq to automate academic research, integrating live arXiv searches with local TF-IDF indexing for real-time paper synthesis. Optimized conversation state management to persist tool outputs, reducing multi-turn latency by over 95% (from 60s to <3s) and ensuring 100% context retention.
End-to-end voice AI pipeline enabling Hindi speech commands to perform natural-language database operations through an LLM-powered agent. Integrated STT, translation, LLM reasoning, and TTS to create a multilingual workflow for account creation, expense logging, and financial data querying.
AI-powered recovery support system featuring a multilingual LLM health assistant, real-time rehabilitation exercise tracking, and automated patient monitoring workflows. Implemented pose-detection based exercise analysis using MediaPipe and OpenCV with automated doctor alerts, AI-generated recovery summaries, and patient activity dashboards.
Production-style RAG pipeline that ingests PDFs, generates MCQs, and stores assessments with adaptive learning engine.
Vision-LLM powered OCR-to-speech workflow with high-accuracy text extraction and seamless voice synthesis.
Fine-tuned DistilBERT model using LoRA for comment classification achieving 87% accuracy with Flask API and Docker containerization.
LangGraph, LangChain, LangSmith, CrewAI, MCP Servers
GPT, Claude, Llama, Mistral, Vision LLMs, RAG, Vector Databases, Prompt Engineering, STT/TTS Pipelines
Python, C/C++, Java, SQL/MySQL β plus 200+ DSA problems solved
React, HTML/CSS/JS, Figma, FastAPI, Flask, Django, Node.js
Docker, GitHub Actions CI/CD, API Deployment, Model Packaging, Microsoft Azure
Data Structures & Algorithms, OOPS, DBMS, Operating Systems
Ownership, Fast Iteration, Bias for Action, Problem Solving, Collaboration
Achieved a SMAPE score under 45, securing 167th rank out of 6,000+ teams in the national competition.
Winner of 3+ hackathons, building production-grade AI and full-stack systems under competitive constraints.