Ameya Jadhav
technical builder (and investor, sometimes)
I study CS (AI/Systems) at Stanford. I work at Neo, OpenAI, Stanford AI Lab, General Catalyst, TreeHacks, Pear Garage, Cardinal Ventures, and more.
more about me
Highlighted Projects
Raspberry Pi Cluster
Custom Rasp Pi Cluster for Scalable Computing
Dave
Devin-esque SWE agent that interacts with local development tools
TestNinja
Autonomous agent platform for robust and context-aware test generation.
PathSense
Revolutionizing indoor mobility with real-time, adaptive AI-enabled guidance.
Sonoverse
Leverages IPFS, NFT Certs, and Ethereum L2 blockchain to secure artists' ownership over their work, automating DMCA claims for small artists.
Custom TCP/IP Implementation
Custom network protocol for optimized communication
Research
HRL for Web Agents
Stanford CS Department
WebHierarch introduces a novel framework that combines large language model reasoning with hierarchical reinforcement learning to advance autonomous web automation. By integrating semantic understanding with experiential skill learning, the system achieves near state-of-the-art performance on complex MiniWoB++ benchmarks, with a 90% average success rate—outperforming both pure RL and LLM approaches. This work establishes a scalable and interpretable foundation for adaptable web agents capable of handling diverse and dynamic online environments.
3D Mesh Reconstruction with Vision Transformers
Stanford CS Department
This work presents a novel framework for reconstructing dynamic 3D meshes from monocular videos using Vision Transformers. The approach integrates D-NeRF scene representations, ViT-based spatial encoding, and Mesh R-CNN decoders, achieving measurable improvements in geometric accuracy and visual fidelity over prior methods. It establishes a proof of concept for applying transformer-based architectures to advance the state of the art in 3D reconstruction
ITP-Enhanced LLM Reasoning
Stanford STAIR Lab (SAIL)
This work introduces a methodology for augmenting mathematical reasoning in LLMs through the incorporation of axiomatic structures from Interactive Theorem Provers. Empirical evaluations reveal statistically significant, albeit modest, enhancement in proof-theoretic capabilities, suggesting this formalism-driven approach offers a promising vector for advancing machine cognition in mathematical domains.
Medical NER and RE from Clinical Narratives
Stanford CS Department
This research develops a computational framework for biomedical entity recognition and relation extraction from clinical narratives. The system identifies semantic relationships between medical concepts, enabling automated reconstruction of patient treatment trajectories and pharmacological response patterns to facilitate evidence-based clinical decision-making.
Digital Threats Against Democracy
Georgia Tech T+ID Lab
This research employs computational linguistic methods to analyze cross-community discourse patterns within Twitter social networks, coupled with human-computer interaction studies examining cognitive load optimization for multi-stream social media monitoring in near-synchronous environments.