Aniruddh Sriram
| Aniruddh Sriram | |
| Occupation | Software engineer, entrepreneur |
|---|---|
| Known for | Co-founder and CTO of Playgent |
| Education | Bachelor of Science in Computer Science and Mathematics, University of Texas at Austin |
Aniruddh Sriram is a software engineer, natural language processing (NLP) researcher, and entrepreneur based in San Francisco, California. He is the co-founder and CTO of Playgent, a company that builds sandbox infrastructure for AI agents. Playgent is a member of the Y Combinator Summer 2025 batch.[1]
Early life and education
Sriram studied computer science and mathematics at the University of Texas at Austin, where he was a participant in the Turing Scholars program. During his time at UT Austin, he conducted research in natural language processing and served as a teaching assistant. His academic research focused on making language models more reliable and factual, and his work has been cited over 260 times according to Google Scholar.[2]
Career
Prior to founding Playgent, Sriram held several positions in the technology industry. He interned at Bloomberg L.P. and DevRev, working on projects related to AI and data frameworks. He also participated in the Wolfram Summer School, where he developed models for disease spread. He subsequently worked as a software engineer at The Voleon Group, a quantitative investment firm, where he focused on securities engineering from July 2024 to June 2025.
Sriram's earlier entrepreneurial experience included co-founding Cornerstone Logistics, where he served as CTO and built technology infrastructure for the SaaS startup.
In 2025, Sriram co-founded Playgent with Neer Jain.[3] The company provides infrastructure for creating high-fidelity sandbox environments in which AI agents can be tested, debugged, and improved. Playgent's environments allow engineering teams to reproduce production failures in controlled settings by recreating the conditions that caused an agent to fail, including user state, data, tool calls, and context. The environments are packaged as a single MCP URL, with tool calls mocked, authentication abstracted, and test data initialized via natural language or JSON specifications. Runs can be exported as OpenTelemetry traces for integration with observability tools or reinforcement learning pipelines.
Playgent has applied its technology to the financial services sector, building reinforcement learning environments for finance. These environments simulate realistic market conditions, document processing workflows, and decision-making scenarios for tasks such as leveraged buyout analysis and take-private transactions. The company curates financial tasks through a network of expert contributors and designs verification rubrics for agent evaluation.
The company is based in San Francisco and is classified in the SaaS, infrastructure, and AI sectors.
References
- ↑ "Playgent: Sandboxes for AI agents". 'Y Combinator}'. Retrieved 2026-03-19.
- ↑ "Aniruddh Sriram". 'Google Scholar}'. Retrieved 2026-03-19.
- ↑ "Playgent". 'Playgent}'. Retrieved 2026-03-19.