Fellowship that makes scientists & deep tech startups use AI like frontier labs do. AI changed software engineering dramatically, science is next.
Joined May 2026
- Tweets65
- Following57
- Followers518
- Likes79
Pinned Tweet
We support the ambitious scientists and deep tech teams who want to move the world forward.
AI multiplies the individuals behind that change
Apply now:
reactorfield.com/apply
xitter.twitterapi.xyz/jamessinka/status/2102…
Cohort 1 represents top talent from industries including mRNA, brain-computer interfaces, longevity, nuclear fusion, and beyond.
This exceptional group uses AI to accelerate their scientific R&D, and make the world a better place.
Deep dive below 🧵:
Excited to introduce @ReactorfieldAI Cohort 1:
@merge - brain-computer interfaces with no implants, $250M seed led by OpenAI and Bain
@episteme - a new R&D institution, cofounded by Louis Andre and Sam Altman
@RetroBio_ - unicorn longevity startup, adding 10 healthy years to human life. In phase 1 trials
@NorthwellHealth - NY's largest healthcare system, with Chief Biotech Officer @itchdoctor
@StrandTx - programmable mRNA medicines out of MIT, lead cancer drug already in patients
@MayoClinic - turning leading clinical science into new ventures represented by @vera_mucaj
@xiliotx - publicly traded biotech making cancer therapies to activate inside tumors
@marathonfusion - producing the world's most valuable isotopes from compact fusion reactors
@KopraBio - viruses that turn cancer cells into factories for their own destruction
@radifymetals - hydrogen-plasma reactors that turn rare-earth oxides directly into metal
@FoundationManf - rapid domestic CNC machining for frontier hardware
@MassMagnetics - US-made high-performance magnets from recycled rare earths
@algabiosciences - turning cattle methane into calories, peer-reviewed cuts of up to 63%
@lumehealth - building a wearable that measures cortisol continuously from sweat
@liu_liu_lemon - smart materials for med tech and robotics, @Harvard PhD, Fulbright Fellow
Alga Biosciences (@algabiosciences) is turning enteric methane into calories for livestock. The YC-backed company has raised roughly $8M and was awarded a USDA Climate Smart Commodities grant. In controlled cattle trials, its Alga 1.0 formulation showed peer-reviewed methane-yield reductions of 39% and 63% at the studied inclusion rates.
1
223
Choosing between Claude and GPT for a biology task starts with knowing where each fails. @LatchBio builds benchmarks that test AI agents on real biological data, making those failure modes visible so scientists can make more informed choices about models and prompts.
Arjun Banerjee (@arjunomics) is joining Reactorfield for a hands-on introduction to benchmarking for scientists: when you need one, how to build your first, and how to test whether a change to your workflow actually improves results.
Scientific workflows often involve unpublished results, proprietary methods, and sensitive data. @getopenwork lets teams build AI workflows with self-hosted open-source models on infrastructure they control. They can then share AI skills and tool connections across the lab, so colleagues can reuse those workflows.
Ben Shafii (@benjaminshafii ), founder of OpenWork, is joining Reactorfield to explore how scientists and deep tech teams can bring AI agents into everyday research, with a focus on handling confidential data and building workflows the whole team can use.
Juan Benet (@juanbenet) looked at YC and saw a model for accelerating science. He began turning Protocol Labs into a network of independent teams, connecting scientists making discoveries with the builders who could turn them into useful tools and products.
We’re welcoming Juan, founder of @protocollabs, to Reactorfield for a 30-minute fireside on how to build an environment where science moves faster.
We’ll explore how to choose problems that unblock other researchers, what teams gain from working together, and where AI can help turn promising ideas into work others can build on.
“You become the PI of your research group of agents.”
Vamsi Varanasi (@feynmanpoint ), cofounder of @TuvaAI, is building that vision.
Tuva builds shared research memory across all of your agents, meaning they can be proactive and contextual about the sittings of your project. Think swarms of agents chasing down open threads, constantly scanning the literature, surfacing ideas, and bulletproofing your results, all while you direct and validate.
We’re happy to welcome Vamsi to Reactorfield for a session on building a company brain for research.
Company brains are a growing requirement in the AI age.
Imagine a promising result in a slide deck and a failed replication in a lab notebook. Any of your AI agents planning the next experiment needs to understand both. Company brains make this possible.
Conor Brennan-Burke (@contextconor) is building @hyperspell around a company brain that connects information across tools and tracks what changed. For research teams, that could mean tracing conflicting results back to the experimental conditions, analysis code, and decisions behind them.
Conor will show Reactorfield's deep-tech founders how to engineer a company brain, exploring what shared, continuously updated context could enable for their research teams.
Every experiment should inform the next. @katyenko and @derekalia are building @muni_bio: a platform for agents to design hundreds of drug candidates, run wet-lab validation, and use that knowledge for future discoveries.
Kat and Derek are joining Reactorfield to show our scientists how to connect autoresearch with real-world experiments.
Application for AI models have transitioned through chatbot era and software engineering, next paradigm is scientific discovery.
Sreeram Kannan (@sreeramkannan) and Soubhik Deb (@soubhikdeb) are taking that idea further with @yukonresearch: an open network where humans and agents explore different approaches, verify improvements against a defined objective for a frontier scientific problem, and build on each other’s best results.
Reactorfield is hosting Sreeram and Soubhik for a session on autoresearch, open science, and how better verifiers can help AI tackle scientific problems.
Reactorfield Cohort 1 starts this week. Excited to announce our Week 1 speakers:
- Becoming an AI-Native Scientist with @DeryaTR_ from @jacksonlab
- AI for Simulating Life Science with @CorinWagen from @RowanSci
- Using AI as an MLE with @eshanchordia from @impulseai_
- From Literature Search to Scientific Writing with @_ayushgarg_ from @answerthisio
Our SF launch at was an absolute hit.
We saw participation across the stack, even some AI lab folks wanting to get a leg up in the world of atoms!
AI for science is heating up.
Thanks to @PresidioBitcoin for the venue. And thank you to everyone who showed up for science!
Reactorfield Cohort 1 applications are closed. We will still review late applications.
We were struck by both the quality and volume of interested applicants. They span scientists from leading institutions and founders building late-stage deep-tech startups.
It made one thing clear: AI is ready for science, and many are excited to wield it. We’re energized to accelerate what comes next!
Excited to be running it back in SF tonight luma.com/hxsbz3xb
Reminder, 3 days left to apply: reactorfield.com/apply
We launched in Boston last night at @enginexyz .
Thank you to every scientist, founder, and builder who came out.
Nobody is asking if AI is good enough for science anymore. People are all digging into workflows: run agents on your own research, build an AI native lab, starting today.
Reactorfield exists to make that happen. Cohort 01 applications are open!
We launched in Boston last night at @enginexyz .
Thank you to every scientist, founder, and builder who came out.
Nobody is asking if AI is good enough for science anymore. People are all digging into workflows: run agents on your own research, build an AI native lab, starting today.
Reactorfield exists to make that happen. Cohort 01 applications are open!
In case you missed us & are in silicon valley next week, come hang out with the team & fellow scientists at the Presidio in SF:
luma.com/hxsbz3xb
1
152
"I don't do anything myself anymore. I just tell the agents what to do."
Professor, immunologist, and biomedical engineer @DeryaTR_ told us this in our first meeting.
He predicted the bio x AI singularity in 2005. Now he's living it: building lab software from his phone in minutes, going from 2 GitHub commits to 4,500+ contributions in months.
Derya is joining Reactorfield as a guest speaker to show our scientists what AI-native research looks like.
OpenAI gets it - we need to distribute the best of our tools in the hands of scientists
Reactorfield helps you use them like a frontier lab
Come meet @jamessinka, @berkbuilds, and early Reactorfield fellows IRL in August!
Boston - August 6th - luma.com/9x6u7kk4
SF - August 15th - luma.com/hxsbz3xb
100k views
Thanks for your interest in accelerating science with AI!
AI is changing science
We makes scientists and deep tech startups AI-native
Built by @jamessinka and @berkbuilds to accelerate science with frontier AI
Supported by @AnthropicAI, @Nvidia, @PrimeIntellect & many more