Work About Research Contact

ADITYA AMBADE

Building AI systems that think, speak, and move.

View Work
Aditya Ambade

I build AI that does things.

AI/ML engineer finishing CS degree, obsessed with RAG pipelines, hybrid search, and LLM orchestration. Currently interning at NeuralVedaX.

Built KinetiAI — a multimodal AI physical therapy assistant that combines pose estimation, voice LLM dialogue, and RAG-grounded clinical responses. 2 IEEE papers submitted.

Targeting a funded Masters in AI/Robotics — Japan, New Zealand, Canada, or Germany. Research output is the north star right now.

Currently → NeuralVedaX Internship
Next → Funded Masters 2026

Projects that matter.

Internship — NeuralVedaX

Production Hybrid Search System

Enterprise-grade retrieval pipeline with multi-strategy fusion

BM25 + FAISS with RRF fusion, cross-encoder reranking, and a unified multi-provider LLM abstraction layer (Anthropic, OpenAI, Gemini). Deployed on Azure VM.

BM25 FAISS RRF BAAI/bge-m3 Cross-Encoder FastAPI Docker Azure
In Production
[ SYSTEM DIAGRAM ]

What I reach for.

Daily Drivers

Python FastAPI FAISS RAG Pipelines LangChain HuggingFace Docker Git LLM Orchestration ChromaDB

Currently Learning

fast.ai Karpathy Zero-to-Hero System Design GSAP Advanced Embeddings

Research

Peer-reviewed work at the intersection of AI and healthcare.

KinetiAI: A Multimodal AI-Assisted ACL Rehabilitation System

AI Fusion Conference 2025 Research Paper Under Review

Multimodal AI in Physical Rehabilitation: A Literature Survey

AI Fusion Conference 2025 Literature Survey Under Review

Let's build something.

Open to research internships, AI/ML roles, and collaborations.