Aayam Bansal

Aayam Bansal
Aayambansal.jpg Aayambansal.jpg
Born	(2007-08-14) August 14, 2007 Kolkata, West Bengal, India
🏳️ Nationality	Indian
🏫 Education	University of Illinois Urbana-Champaign (dropped out)
💼 Occupation	Computer Researcher, Entrepreneur
Known for	Co-founder of Synthetic Sciences (YC W26), AI research in reinforcement learning
🌐 Website	www.aayambansal.com

Aayam Bansal (born 14 August 2007) is an Indian computer researcher and entrepreneur, best known as the co-founder and CEO of Synthetic Sciences, a Y Combinator Winter 2026 (YC W26) backed startup building AI-powered scientific research platforms. His work spans reinforcement learning, machine learning, and artificial intelligence, with publications at top-tier venues including NeurIPS, ICML, ICLR, AAAI, and CVPR workshops.

Early Life and Education

Aayam Bansal was born on August 14, 2007, in Kolkata (formerly Calcutta), West Bengal, India. He attended Delhi Public School Ruby Park, Kolkata, where he developed an early interest in computer science and artificial intelligence. After graduating high school in 2025, Bansal enrolled at the University of Illinois Urbana-Champaign but dropped out shortly after to pursue entrepreneurship full-time.

Career

aisock (Acquired)

Before founding Synthetic Sciences, Bansal founded aisock, a healthcare technology company that developed an AI-powered orthopedic sock designed to help correct metatarsus adductus in children. The product was patented in the United States, Singapore, and India, and was acquired for a five-figure sum when Bansal was 17 years old. He also built COVID-19 helpline infrastructure for the Indian government that served over 50,000 people daily at its peak.

Synthetic Sciences (YC W26)

In 2025, Bansal co-founded Synthetic Sciences with Ishaan Gangwani. The company was accepted into Y Combinator's Winter 2026 batch and has raised $1.4 million in pre-seed funding from investors including Charlie Songhurst, Firestreak Ventures, Pareto Holdings, Pioneer Fund, Amplo VC, Noar Ventures, and an Andreessen Horowitz (a16z) scout fund.

Synthetic Sciences operates as an "AI co-scientist" platform, enabling researchers to delegate complex scientific tasks to autonomous AI agents. The platform handles the full research loop: literature review, hypothesis generation, experiment design, code execution, and publication-ready drafting. The company's biology mode achieved state-of-the-art performance (92%) on the BixBench Verified benchmark, outperforming competitors including Claude Code (65.3%) and OpenAI Agents SDK (61.3%).

The company also offers "Flywheel Mode," allowing organizations to train proprietary models using their own research data through supervised fine-tuning (SFT) and reinforcement learning (RL).

Research Career

Bansal has conducted research at several prestigious institutions:

• MIT CSAIL — Kelis Lab, working on foundation models, time-series analysis, and scientific reasoning under Professor Manolis Kellis
• Carnegie Mellon University (CMU) — Research on machine learning and AI systems
• National University of Singapore (NUS) — Machine learning group research
• University of Oxford — Worked with Dr. Ken Kahn on reasoning and scientific agents

Research Contributions

Publications

Bansal has published research at leading AI and machine learning venues:

• "Value Decomposition Fails in Anti-Coordination: A Systematic MARL Comparison for Dynamic Spectrum Access" — Presented at NeurIPS 2025 Workshop on AI4Science (Spotlight). This work presented the largest systematic comparison of multi-agent reinforcement learning methods for dynamic spectrum access, demonstrating that value decomposition methods (VDN, QMIX, QPLEX) dramatically underperform even random baselines in anti-coordination environments.

• "Executable Claims: Turning Manuscript Statements into Verifiable Evidence Capsules at Writing Time" — AAAI 2026 Workshop on AI for Research, February 2026.

• "Grid Stability Analysis with High Renewable Integration: Simulation-Based Assessment and Mitigation Strategies" — IEEE PES/IAS PowerAfrica Conference 2025, May 2025.

• "Advanced Machine Learning Approaches for Formula 1 Race Performance Prediction" — ArXiv Preprint, 2025.

His research interests include reinforcement learning, multi-agent systems, foundation models for science, and AI safety.

Research Impact

Bansal's work on multi-agent reinforcement learning has contributed to understanding the limitations of value decomposition methods in anti-coordination problems, with implications for 6G wireless networks, cognitive radio, and dynamic spectrum access systems.

Recognition and Media Coverage

Bansal has been featured in several international publications:

• New York Post — "Teens Are Launching AI Companies and Making Big Money" (February 2026)
• The Economic Times — "Indian Teens Among Youngest Founders Accepted into Y Combinator W26" (December 2025)
• VnExpress — "Meet Teenage Founders Building Fast-Growing AI Startups"
• The Telegraph — "A Teenager's Journey to Y Combinator"
• Silicon Valley Post — "Top Bio & Health Startups from YC W26"
• Kolkata Calling — "Teenagers from Calcutta and Pune Take InkVell to Y Combinator"

He was also featured on the Hello Kolkata! podcast (Episode #58) discussing his entrepreneurial journey.

Personal Life

Bansal is based in San Francisco, California. He is a Z-Fellow and recipient of an Emergent Ventures grant. He maintains active profiles on academic and professional networks including Google Scholar, ResearchGate, LinkedIn, and GitHub.

External Links

• Official website
• LinkedIn profile
• X (Twitter) profile
• Instagram profile
• GitHub profile
• Google Scholar profile
• ResearchGate profile
• Crunchbase profile
• Synthetic Sciences on Y Combinator

See Also

• Synthetic Sciences
• Y Combinator
• Reinforcement learning
• Multi-agent reinforcement learning
• Artificial intelligence in healthcare
• Scientific computing

---