Backend & AI Engineer

I build systems
that think,

Crafting RAG pipelines, LLM integrations & OCR systems that solve real problems. Based in India, building for the world.

View Projects Get in Touch
2+
Years building
10+
AI projects
About

Engineer by logic,
builder by instinct.

I'm Prithvi — a Backend & AI Engineer who loves turning complex AI research into production systems that actually work. Currently building at Lexlegis.ai, where I work on legal AI pipelines.

I believe the best code is code you don't have to write twice. I'm obsessed with RAG architectures, OCR pipelines, and agentic systems — building things that are fast, reliable, and surprisingly smart.

When I'm not at work, I'm exploring local AI with Ollama, building personal tools, and thinking about how AI will reshape the next decade of software.

🏢
Currently at
Lexlegis.ai
🎓
Education
B.Tech + CDAC
📍
Location
India
⚡
Focus
AI Infrastructure & Backend
🔭
Exploring
Agentic AI & Offline LLMs
Skills

What I work with.

🤖 AI & LLMs
RAG Pipelines Gemini OpenAI API Ollama Vector DBs Agents
🔧 Backend
Python FastAPI REST APIs PostgreSQL Redis Celery Docker
👁️ OCR & Vision
Tesseract Google Vision PDF Parsing Image Processing Marathi OCR
🛠️ Tools & Infra
WhatsApp API Git Linux Nginx GitHub Actions Postman
Projects

Things I've shipped.

⚖️
Production
Legal AI Agent
Agentic legal assistant built on top of RAG pipelines. Processes legal documents, extracts clauses, answers queries with citations.
RAG FastAPIVector DBLLM
🛒
Production
WhatsApp Marathi OCR
Production WhatsApp bot that processes shopping list images and PDFs in Marathi. Uses OCR + Gemini to extract, structure and digitize handwritten lists for local vendors.
WhatsApp APIOCR GeminiPythonMarathi NLP
🤖
↗
Personal Assistant
Personal AI ecosystem integrating mobile, laptop and IoT. Local-first design using Ollama — no cloud dependency. Manages tasks, answers questions, controls smart devices.
OllamaIoT Local LLMPythonAgents
📄
↗
PDF Intelligence Pipeline
Extracts, chunks and semantically searches across large document collections. Built with custom chunking strategies for legal and financial documents with 90%+ retrieval accuracy.
RAGPDF EmbeddingsFAISSFastAPI
🚀
Contact

Let's build something.

Open to freelance projects, collaborations, and interesting problems. I respond to everything — just reach out.

✉️ Email me ⌥ GitHub in LinkedIn