NeoSigma

Let your agents learn from experience in production.

[NeoSigma @neosigma_ai]

The next era of AI engineering is self-improving agentic systems! Really excited to share what we are building at NeoSigma!

Self-maintaining agent systems represent a shift in how we build and operate software. We, at NeoSigma are building the infrastructure to support this feedback loop in real-world systems, helping teams capture failures, convert them into structured evaluation signals, and use them to drive continuous improvements in agent behavior.

[Gauri Gupta @gauri__gupta]

We @neosigmaai @RitvikKapila are building the future of self-improving AI systems! By closing the feedback loop between production data and system improvements, we help teams capture failures, convert them into structured evaluation signals, and use them to drive continuous improvements in agent behavior.

We show how our system works on Tau3 bench across retail, telecom, and airline domains. Agent performance on the validation set (with a fixed underlying model, GPT5.4) improves from 0.56 → 0.78 (~40% jump in accuracy).

[NeoSigma @neosigma_ai]

@gauri__gupta read more here: https://t.co/0Nn6uqXmJo

[Gauri Gupta @gauri__gupta]

Excited to push the frontier of auto-research and RSI @neosigma_ai
directly built for agentic systems to learn from experience in production :)

https://t.co/4Ds77YuouI

[NeoSigma @neosigma_ai]

We're proud that NeoSigma has signed the Open Weights and American AI Leadership letter.
We believe the future of AI shouldn't be built behind closed doors.

Open models and ecosystems have accelerated research, enabled reproducible science, and created a more competitive AI ecosystem. The next breakthroughs won't come from models alone — they'll come from the infrastructure, evaluations, more importantly ecosystems built around them.

https://t.co/WxSGAfimFo

[NeoSigma @neosigma_ai]

If you’re attending @ycombinator SUS and are excited about building the next frontier of agent architecture, we would love to chat!

[Ritvik Kapila @ritvikkapila]

If you're attending or in town for @ycombinator SUS and we didn't get a chance to meet, we @neosigma_ai would love to talk. Please apply here - https://t.co/FPDTXzUUw6

We are building our founding team, and are an intensely technical team of builders and ex-founders pushing the frontier of agentic systems.

[NeoSigma @neosigma_ai]

come build with us!

[Gauri Gupta @gauri__gupta]

We’re building our founding team @neosigma_ai.
If you are impacted by layoffs at @Amazon AGI, we would love to chat!

@neosigma_ai is a product-driven research lab building infrastructure for agents to learn from experience in production. We are an intensely technical team of researchers, engineers, and previous founders pushing the frontier of agentic systems and redefining the interface between humans and AI.

We're hiring for roles across research, engineering, and infrastructure. If you are excited about our mission, we would love to chat!

[NeoSigma @neosigma_ai]

Great day of research and building sessions with the @agihouse_org community — excited to be part of the push on autoresearch, idea diversity, and research taste.

[Gauri Gupta @gauri__gupta]

It was great hosting an incredible day of research and build sessions on auto research along with @MillionInt and @neosigma_ai

[muhashmi @muhashmii]

Recommended reading, went down the autoresearch rabbit hole a few months ago myself.

We @daytonaio will be speaking about our role in this and reinforcement learning overall at AGI House this Saturday, alongside @neosigma_ai, @ExaAILabs, and @CoreAutoAI. Come chat with us!

[NeoSigma @neosigma_ai]

@shagunuppls building a real high fidelity environment is the pre-cursor to a great policy.

[NeoSigma @neosigma_ai]

https://t.co/PDwGDe6d0y

[AGI House @agihouse_org]

AI agents are automating research end to end. The hard part is trust.

Ahead of our Autoresearch Build Day, here are eight essential reads shaping autoresearch. 1/11 🧵

[Daytona @daytonaio]

"An agent is only as capable as the environment it can safely act in."

@neosigma_ai published a great deep dive on building agent workspaces on Daytona. Check it out

👇

[Gauri Gupta @gauri__gupta]

An agent is only as capable as the environment it acts in. Here's how we built our secure, scalable workspaces @neosigma_ai where agents can execute, explore, and learn from experience.

[Ritvik Kapila @ritvikkapila]

How we built production execution environments for agents, pushing the pareto frontier between control, execution, security and data planes!!

[Abhiram2k03 @Abhiram2k03]

We built around four constraints that had to coexist: low-latency execution, real machine environments (not shells), strong isolation, and fully reproducible runs. Most infrastructure compromises on one or more. We designed ours not to.

[Gauri Gupta @gauri__gupta]

Private evals is all you need!

[Gauri Gupta @gauri__gupta]

Come build with us at AGI house this Sat!

Speaking alongside @MillionInt and @pushmeet about what we are building @neosigma_ai!

We will be talking about building auto-research and building autonomous agents that learn from real-world experience.

[NeoSigma @neosigma_ai]

Excited to partner with @agihouse_org, @ExaAILabs, and @daytonaio for the Auto-Research Summit & Build Session on July 18.

Looking forward to bringing together researchers, builders, and engineers to explore around autonomous improving agents and the infrastructure that powers them.

See you there!

[AGI House @agihouse_org]

Auto-Research Summit & Build Session is coming to AGI House on July 18, in partnership with @ExaAILabs, @neosigma_ai and @daytonaio.

Featuring speakers from @CoreAutoAI, @GoogleDeepMind, Google Cloud, Daytona, Google, and NeoSigma. RSVP now ⬇️ https://t.co/oGFWYoLj6v

[NeoSigma @neosigma_ai]

We at NeoSigma are hosting an Auto Research Summit and Build Session at @agihouse_org on July 18, 2026, in partnership with @daytonaio and @ExaAILabs

Joining us with keynotes
- Gauri Gupta and Ritvik Kapila, Founders, @neosigma_ai
- Jerry Tworek — Co-founder, @CoreAutoAI , ex - RL Lead @OpenAI
- Pushmeet Kohli — VP, @GoogleDeepMind & Chief Scientist, @googlecloud
- Muhammad Hashmi - building computers for agents @daytonaio
- Ivan Nardini, Senior AI Dev Engineer @Google

A full day of research talks and hands-on building around autonomous improving agents and the infrastructure that powers them.

We're also hiring — if you're curious about powering agents that learn from experience in production, come find us.

Doors open at 10 AM. See you there!
@gauri__gupta @ritvikkapila @MillionInt @pushmeet @muhashmii @ivnardini

Apply: https://t.co/5mgSYGd9s9

[NeoSigma @neosigma_ai]

We're hiring an Infrastructure Engineer at NeoSigma.

At NeoSigma, we're building the infrastructure that enables AI agents to learn from experience. You'll own and scale the cloud foundation behind our platform—from cloud infrastructure and Kubernetes to deployment automation, security, and enterprise-grade deployments for regulated customers.

We're looking for someone with 1-2+ years shipping real infra at a startup, strong fundamentals in cloud (GCP/AWS/Azure), Kubernetes, and Terraform, experience with enterprise security posture for deployments, and a high-ownership mindset.

If you enjoy building reliable systems that power cutting-edge AI, we'd love to talk.
More details and apply here: https://t.co/Ky2HrE96GQ