AI Engineering Undergraduate @ IIT Madras | Kaggle Expert (Top 4%)
I am a final-year student specializing in building Agentic RAG workflows and scalable End-to-End ML systems. I bridge the gap between complex Generative AI reasoning and high-performance backend infrastructure.
I have a strong foundation in Data Structures & Algorithms and a passion for reducing inference latency in AI applications.
- 🔭 I’m currently building Multi-Agent Systems using LangGraph and FastAPI.
- 👯 I’m looking to collaborate on Generative AI and Backend Optimization projects.
- 💬 Ask me about: LLMs, Vector Databases, RAG Pipelines, and System Design.
- 🏆 Achievements: Kaggle Expert (Ranked 341 Global) and 200+ LeetCode Problems Solved.
LangGraph · FastAPI · Gemini · Redis · Docker · Vue · Chrome Extension
- Built a multi-agent hiring simulator with 3 personas (85% eval accuracy)
- Reduced application time 45 min → <2 min
- Cut TTFT by 90% using async pipelines & semantic caching
- Full observability with LangSmith tracing
👉 This is not a demo. This is an engineered system.
LangGraph · FAISS · FastAPI · Async Streaming
- Sub-500ms TTFT via real-time streaming
- Persistent multi-turn conversations over 1h+ transcripts
- MMR search eliminated 90% redundant context
Multi-Agent Vision + Weather + Agronomy System
- Image-based crop diagnosis grounded with real-time weather
- Strict schema enforcement → zero parsing failures
- End-to-end reasoning in <3s latency
XGBoost · MLflow · Streamlit
- Benchmarked 4 ensemble models on 100k+ records
- Fully reproducible training with MLflow
- Deployed for non-technical users
- 🥇 Kaggle Expert — Top 4% globally (1 Silver, 9 Bronze)
- 🧠 LeetCode — 200+ problems solved (100+ Medium)
- ⭐ HackerRank — 5★ Gold (SQL)
- 🧑💻 GitHub: https://github.com/GyanPrakashKushwaha
- 💼 LinkedIn: https://linkedin.com/in/gyanprakashkushwaha
- 📊 Kaggle: https://kaggle.com/gyanprakashkushwaha
If you’re here to skim — fine.
If you’re here to collaborate or hire — even better.