Utsav Singhal profile photo

Utsav Singhal

Co-Founder and CTO at Interact AI (since February 2025, present). Also serves as an Advisor at GOAT LIFE (since September 2023, present).

AI Agent OrchestrationAI Infrastructure & ScalingWeb EngineeringCloud SecurityFounder/Executive LeadershipProduct Development

About

I'm Utsav Singhal, Co-Founder and CTO at Interact AI and an Advisor at GOAT LIFE. My career has taken me from research roles at the University of Tokyo and IBM to co-founding multiple startups, including CaptainTesseract and GOAT LIFE, which raised significant funding on Shark Tank India. I'm passionate about building practical AI applications that handle real-world, messy data, and I'm deeply engaged in AI agent orchestration, infrastructure scaling, and the founder ecosystem. I offer insights on cutting-edge AI development, founder advice, and tools like InteractPitch to help with fundraising. I'm always looking to connect with fellow founders, AI professionals, and anyone interested in pushing the boundaries of what AI can do in practical settings.

Networking

What I can offer

  • Insights on cutting-edge AI development
  • Practical founder advice
  • Critical analysis of AI trends
  • A tool (InteractPitch) to help founders with fundraising
  • Hosting and participating in networking events to facilitate community connection

Looking for

  • expanding my professional network
  • exploring mutual opportunities in AI and startup ecosystem

Best fit for

Co-foundersStartup foundersAI product designersMarketersAI infrastructure professionalsThe broader San Francisco tech/AI community

Current Interests

AI Agents & OrchestrationAI Infrastructure & ScalingAI for Business & MarketingFounder EcosystemAI Industry Dynamics

Background

Career

Started as Undergraduate Research Fellow at The University of Tokyo (2019), then Research Intern at IBM (2020), Co-Founder at CaptainTesseract (2020-2021), Engineering Intern at Eightfold (2021), Member of Technical Staff at Eightfold (2022-2023), Software Engineer at Eightfold (2023-2024), Advisor at GOAT LIFE (2023-present), Co-Founder & CTO at Interact AI (2025-present).

Education

Dual Degree (Btech + Mtech) in Mathematics and Computing from the Indian Institute of Technology, Delhi (2017 - 2022). Served as Captain of the Basketball Team for Udaigiri House (2018, 2019).

Achievements

  • At IBM (2020): Designed a new security model for Order Preserving Encryption that is 'strictly stronger than OPE' and reduces decryption accuracy from ~100% to hiding the order, with only a logarithmic increase in client-side storage.
  • At University of Tokyo (2019): Co-developed the Diamond Matrix Powers Kernel (DMPK), which reduced inter-node processor communication 'to almost half.' The work was accepted at HPCAsia 2020.
  • At GOAT LIFE: Co-founded the company which later raised INR 2 Crore from Aman Gupta and Anupam Mittal on Shark Tank India.
  • At Interact AI: Launched 'InteractPitch,' which became the #1 Product of the Day on Product Hunt. The launch drove thousands of global account creations, necessitating a quadrupling of AWS ECS pods to handle traffic.
  • Holds a patent (filed during IBM internship, discovered via Perplexity Patents).

Opinions

  • "AGI is not there yet." Current powerful tools are "impressive for demos and structured tasks" but fail at "real-world, unstructured, high-volume messy data." The path forward is "building more crazy, real-world applications" to expose gaps.
  • "Prompt injections will break your agent before your model gets better." Argues that filters are insufficient and agents must be designed like operating systems with isolation, explicit user confirmation, and logging.
  • Is highly critical of the current AI optimization (GEO/AEO) landscape, calling much of it a "SCAM" comprised of "smoke and mirrors." Believes victory comes not from chasing rankings but from "rewriting the algorithms that define them" using real-time browser agents and live prompt intelligence.
  • Argues that "Daily experience is the real moat," not benchmarks. Praises OpenAI's internal "code red" on quality as a correct shift from feature-shipping to fixing the core user experience.
  • Asserts that "Single region is a single point of failure. Design for failure, then prove it on game day."