Sachin Dharashivkar

Sachin Dharashivkar is the Founder and CEO of AthenaAgent, which trains specialised coding agents for complex enterprise codebases. 

With over a decade of experience training models with reinforcement learning, he developed algorithmic trading agents at JP Morgan, trained self-driving car agents at Huawei Research, and co-trained Aryabhata 1.0, a 7-billion-parameter reasoning model that outperformed frontier models from OpenAI and Google on JEE math. 

At AthenaAgent, he is also developing Savante, a set of coding agents designed for complex enterprise codebases. These agents are focused on tasks such as deep code understanding, architectural refactoring, feature development, and intelligent test generation, particularly for high-coupling enterprise systems where conventional coding copilots often struggle.

Before founding AthenaAgent, Dharashivkar worked in JPMorgan Chase & Co.’s AI Acceleration group, where he built low-latency stock prediction models operating in 10–100 millisecond windows, contributed to trading infrastructure systems, and designed deep reinforcement learning agents for high-volume equity execution.

He also previously worked at Unity Technologies, where he trained deep learning agents for multiplayer gaming environments, including an AI agent for the collaborative game Overcooked by distilling slow planning systems into fast real-time neural networks.

He holds an MS in Computer Science from the University of Massachusetts, Amherst.