I'm an AI engineer working on deep learning for autonomous driving and on-device LLMs.
Most of my work focuses on building efficient, privacy-first AI systems that run entirely offline.
-
local-llms-on-android
Run Qwen2.5, Qwen3, Gemma, and LLaMA models offline on Android with streaming token generation, KV-cache reuse, and multi-turn chat -
multi-task-neural-networks-for-ADAS
multi-task neural network architecture -
llm-lab-from-scratch-to-fine-tuning
Tools for training and fine-tuning large language models from scratch.
- email - dinesh.soudagar@gmail.com