CS undergrad — AI/ML honors, CGPA 9.5 at IPS Academy, Indore — graduating June 2026.

At WeDecode.io I build the backend for gym environments that train and evaluate AI agents: synthetic data generation with LLMs, RL-based evaluation pipelines, and open-source SaaS stacks (OpenEMR, Frappe, Rocket.Chat) wired into controllable, production-like worlds an agent can actually operate inside.

Before that — an OCR + blockchain document-verification pipeline at THDC India, and an LLM lead-gen system at Valency Renewable (8K+ records, ~35% efficiency lift).

I shoot photographs and write the rest of the time. The portfolio carries those.

🧠 Recall — AI knowledge-management platform

Semantic RAG over six content types, a conversational memory assistant, real-time handwritten-math recognition with step-by-step solving, and a slide-deck generator from a single topic. React 18 · TypeScript · Supabase · pgvector · Gemini 2.5 · GPT-5 · KaTeX · Deno

🎙️ Anuvaad AI — multilingual dubbing studio

Speech recognition → neural translation → voice cloning across 50+ languages. Ships an article-to-podcast converter alongside it. Python · Flask · Whisper · Coqui-TTS · ElevenLabs · LangChain

📋 AcadBoost — academic profiles + AI résumé engine

Role-based records with automated LaTeX résumé generation (94% ATS compliance) and an analyzer that tells you why a résumé fails ATS, not just that it does. Python · Streamlit · Gemini · MongoDB · LaTeX · LangChain

stack = {
    "languages":  ["Python", "C++", "SQL"],
    "ml_ai":      ["PyTorch", "TensorFlow", "Scikit-learn", "LLMs", "RAG", "RL"],
    "frameworks": ["FastAPI", "Flask", "LangChain", "Hugging Face", "Django"],
    "data":       ["PostgreSQL", "MongoDB", "Redis", "Supabase", "pgvector"],
    "infra":      ["Docker", "GitHub Actions", "Google Cloud", "Git"],
    "tools":      ["Pandas", "NumPy", "Selenium", "Power BI", "Postman"],
}

_{📍 22.72°N 75.86°E · Indore, India  ·  open to roles from July 2026}