Software engineer and researcher focused on Agentic AI, LLM systems, and applied ML. Currently at MANIT Bhopal pursuing M.Tech in CSE, with research centered on advanced time-series learning frameworks. I work across the full stack — from low-level C++ systems to production LLM pipelines.
Core interests: autonomous agent architectures, RAG systems, transformer-based research, and building ML systems that ship.
Orchestration LangChain · LangGraph · AutoGen · CrewAI
Retrieval RAG · FAISS · ChromaDB · Pinecone · Embeddings
LLM Providers OpenAI API · Hugging Face · Llama · Mistral
Serving FastAPI · Streamlit · REST · WebSockets
Observability LangSmith · Prompt Engineering · Evals
|
GNSS Anomaly Detector
High-performance anomaly detection for geodetic GNSS time-series data. Outperforms CNN/LSTM baselines with custom feature engineering and ensemble methods.
|
AI Receptionist — HIL System
Human-in-the-loop conversational AI with persistent memory, real-time STT, and LangChain-powered context tracking. Deployed via FastAPI.
|
|
Drone Navigation via RL
Autonomous navigation with hybrid PPO + A* algorithm in a custom OpenAI Gym environment. Multi-objective reward shaping for efficiency and safety.
|
Advanced Algo Bot
Low-latency C++ trading bot ingesting live market data via libcurl and jsoncpp. Real-time signal generation with technical indicator computation.
|
|
Qwen3.5-9B Benchmark Suite
Comprehensive evaluation framework for Qwen3.5-9B across reasoning, math (GSM8K), coding (HumanEval, MBPP), safety (TruthfulQA), and long-context retrieval.
|
Sorting Algorithm Visualizer
Real-time interactive visualizer for classic sorting algorithms — Bubble, Quick, Merge, Heap, Insertion, Selection — built with C++ and SDL2.
|
2 IEEE Publications — computational intelligence, sensor fusion, and time-series ML. Work presented and published through peer-reviewed IEEE conferences.
Engineering intelligence — where clean code meets smart data.