I'm Atharva
Building world’s fastest & smartest computer agents
AG

About

Neuro symbolic frameworks for LLM scaling (test and train) | building browser agents at AGI

Work Experience

A

AGI, Inc.Trained a computer-use agent ranked #1 on OSWorld & AndroidWorld; superhuman on both. RealEvals.xyz: Realistic mini-internet for training & evaluating web agents (benchmark + RL envs). Used internally by leading AI labs to train their models.

May 2025 - Present
Founding Researcher
Trained a computer-use agent ranked #1 on OSWorld & AndroidWorld; superhuman on both. RealEvals.xyz: Realistic mini-internet for training & evaluating web agents (benchmark + RL envs). Used internally by leading AI labs to train their models.
A

Arizona State UniversityImproving reasoning capacities of LLMs through neuro-symbolic frameworks. Leveraging foundational models as a general solver for complete information environments. Interactive uncertainty reduction for efficient vision-language spatiotemporal navigation.

Jan 2024 - Aug 2025
Research Assistant
Improving reasoning capacities of LLMs through neuro-symbolic frameworks. Leveraging foundational models as a general solver for complete information environments. Interactive uncertainty reduction for efficient vision-language spatiotemporal navigation.
S

Stanford UniversityEarly heart deterioration detection in pediatric cardiovascular patients.

Oct 2023 - Aug 2025
Lead Researcher
Early heart deterioration detection in pediatric cardiovascular patients.
S

Samsung ElectronicsLed a team of interns in developing a federated shot suggestion system for the camera suite of Samsung phones, delivering production-ready models for over 30 scenes.

Jan 2022 - Oct 2022
Researcher
Led a team of interns in developing a federated shot suggestion system for the camera suite of Samsung phones, delivering production-ready models for over 30 scenes.
Research Publications

Papers & Patents

  • S

    Stop Anthropomorphizing Intermediate Tokens as Reasoning/Thinking Traces!

    arXiv preprint

    S Kambhampati, K Stechly, K Valmeekam, L Saldyt, S Bhambri, V Palod, ... - arXiv preprint arXiv:2504.09762
  • R

    REAL: Benchmarking Autonomous Agents on Deterministic Simulations of Real Websites

    arXiv preprint

    D Garg, S VanWeelden, D Caples, A Draguns, N Ravi, P Putta, N Garg, ... - arXiv preprint arXiv:2504.11543
  • A

    A systematic evaluation of the planning and scheduling abilities of the reasoning model o1

    Transactions on Machine Learning Research

    K Valmeekam, K Stechly, A Gundawar, S Kambhampati - Transactions on Machine Learning Research
  • B

    Beyond semantics: The unreasonable effectiveness of reasonless intermediate tokens

    arXiv preprint

    K Stechly, K Valmeekam, A Gundawar, V Palod, S Kambhampati - arXiv preprint arXiv:2505.13775
  • R

    Robust planning with llm-modulo framework: Case study in travel planning

    arXiv preprint

    A Gundawar, M Verma, L Guan, K Valmeekam, S Bhambri, ... - arXiv preprint arXiv:2405.20625
  • P

    Planning in Strawberry Fields: Evaluating and Improving the Planning and Scheduling Capabilities of LRM o1

    arXiv preprint

    K Valmeekam, K Stechly, A Gundawar, S Kambhampati - arXiv preprint arXiv:2410.02162
  • S

    Superior Computer Chess with Model Predictive Control, Reinforcement Learning, and Rollout

    arXiv preprint

    A Gundawar, Y Li, D Bertsekas - arXiv preprint arXiv:2409.06477
  • S

    SQL injection and its detection using machine learning algorithms and BERT

    International Conference on Cognitive Computing and Cyber Physical Systems

    S Lodha, A Gundawar - International Conference on Cognitive Computing and Cyber Physical Systems, 3-16
Contact

Get in Touch

Want to chat? Feel free to reach out to me via email, connect with me on LinkedIn, or follow me on X/Twitter. I'm always open to discussing research opportunities and collaborations.