
Killing Death w Autonomous labs @cogworksai | Emergent Ventures Grantee | Stanford ASES fellow | lightbearer @ https://teapawt.net/t.co/Dw7EQss6Wg
Joined April 2022
- Tweets1K
- Following1.5K
- Followers1.8K
- Likes1K
Dhruv Agarwal retweeted
Introducing FrontierML, a benchmark for evaluating AI agents on autonomous empirical ML R&D.
Most existing benchmarks for ML "research" give the agent a validation set or verifier that closely tracks its final score. In practice, human researchers rarely have that. Much of empirical research is deciding how to measure progress on a problem and how far to trust those measurements.
In FrontierML's 25 open-ended problems, agents have to operationalize progress through their own experiments, constructing local proxies and calibrating how far to trust them, while working toward an objective they never directly observe. Their submitted model or method is then scored on a held-out evaluation.
Dhruv Agarwal retweeted
Replying to @archgen_
@archgen_ is building skyscrapers for semiconductor chips
Shrinking transistors is getting harder. Stacking them makes heat harder to manage. We’re building the tools to design chips in three dimensions.
We stack transistors vertically, shorten connections, and optimize timing and heat together.
Our foundational model for thermal simulation, predicts temperature roughly 5,000× faster than conventional EDA tools.
Our placement engine ranked #1 in HRT’s macro-placement challenge
We’re working with some of the largest fabs in the world to tape out our first chip in April 2027
archgenlabs.com
Dhruv Agarwal retweeted
Had a really thoughtful conversation with Greta Sneideriene from TTP plc
at @hupo_org
Greta works right at the intersection of biochemistry, assay development, and life-science instrumentation, translating molecular biology workflows onto automated systems, from low-volume liquid handling to sequencing and other high-throughput workflows.
We ended up going deep on a problem we think about constantly at @cogworksai:
How do you take a biological workflow that works manually, make it truly automation-ready, preserve assay integrity, connect fragmented instruments and data systems, and then layer AI on top without creating more complexity?
We talked about assay integration, instrumentation bottlenecks, high-throughput biology, data fragmentation, and where AI agents can realistically sit on top of existing lab equipment.
@furst_fly #HUPO2026
Dhruv Agarwal retweeted
Found another proteomics scientist interested in trying @cogworksai
Met Prof. Golfam Ghafourifar from the University of the Fraser Valley, Canada at @hupo_org . Her work on automated proteomic workflows + rapid protein digestion is super aligned with what we’re building.
Excited to test this together <3 @furst_fly #HUPO2026
Dhruv Agarwal retweeted
Had a great conversation with Uli Ohmayer from NEOsphere at
@hupo_org. NEOsphere Biotechnologies(based in Germany) is essentially a high-throughput proteomics company for drug discovery, especially targeted protein degradation.
For @cogworksai, the especially interesting part is that they describe their platform as fully automated, high-throughput proteomics spanning cell treatment through MS and data interpretation. That makes them a good company to learn from about how an industrial proteomics workflow is actually automated at scale.
Always useful to compare notes with people building directly inside the proteomics ecosystem.
@furst_fly #HUPO2026
Dhruv Agarwal retweeted
Opentrons @opentrons x Cogworks @cogworksai
Great running into the Opentrons team at HUPO @hupo_org, got a closer look at the Flex and its multichannel pipetting setup.
Exciting that it is not just liquid handling, but how platforms like this can become building blocks for programmable, agent-driven lab workflows.
The path from lab automation → autonomous experimentation is getting very real. 🧪🤖
@furst_fly #HUPO2026
Dhruv Agarwal retweeted
Had a great technical deep dive with the Bio-Rad @BioRad team on APIs vs SDKs, instrument integration, and what lab automation actually looks like in practice. @hupo_org
One thing that really stuck with me: scientists still routinely use pen drives to move instrument readouts around 😭
And honestly, it makes sense. Scientific instruments have been designed around workflows scientists already know touchscreens, manually moving plates in and out, reading results on the instrument display, exporting files to USB drives.
But automation changes the interface completely.
Instead of:
scientist → touchscreen → plate transfer → instrument → USB drive → analysis
we’re heading toward:
sample → instrument → API → cloud → analysis → next experimental action
The interesting challenge isn’t just making instruments “automatable.” It’s designing the transition from workflows scientists are deeply used to, toward labs where instruments, software, and data systems can talk to each other automatically.
That gap between how labs work today and where autonomous labs @cogworksai are headed is fascinating.
@furst_fly
#HUPO2026
Dhruv Agarwal retweeted
Tecan Japan 🇯🇵 @Tecan_Talk at HUPO 2026: a global provider of automated laboratory instruments and solutions for life sciences, clinical diagnostics, and biopharmaceutical applications.
At Cogworks @cogworksai , we do partnerships with lab automation vendors to drive faster close loop experients.
@furst_fly
#HUPO2026 #labautomation
Dhruv Agarwal retweeted
got a Japanese client who listens to my fav nihongo artist 😭<3
maybe this is what product-market fit feels like!!! @hupo_org
@cogworksai @furst_fly
#HUPO2026 #labautomation
RT @b1shtream: Met the Quantum-Si @Quantum_Si at HUPO 2026, really interesting to think about them not just as a protein sequencing platfor…
Dhruv Agarwal retweeted
Found a cool industry advisor from Australia at @hupo_org 🇦🇺. Some conversations just click. Prof. Vera Ignjatovic AM, a prominent Australian medical researcher, innovator, and former Olympic athlete known for her work in paediatric haematology and proteomics.
Great discussions on finding your first customer, GTM, the difference between customers and end users, selling into academic labs vs enterprises, proteomics entrepreneurship, and turning great science into sustainable companies. @cogworksai @furst_fly
#HUPO2026
Dhruv Agarwal retweeted
Let's chat about solving the bottlenecks in biology, wetlabs and lab automation at HUPO @hupo_org 2026: The Human Proteome Organization World Congress tomorrow! We'll be showing our demo videos of few proteomics assays we automated at Cogworks.
@b1shtream @furst_fly
#hupo2026
We’re excited to attend HUPO 2026: the Human Proteome Organization World Congress in Singapore, taking place September 27–October 1.
We’re looking forward to learning about the latest advances in proteomics technologies, biomarker discovery, multi-omics, AI, and precision medicine and connecting with researchers, clinicians, and industry leaders shaping the future of proteomics.
If you’re attending, we’d be happy to connect and chat about automating high-throughput workflows across proteomics and how AI-driven automation can help accelerate research and scale laboratory operations @cogworksai.
Venue: Suntec Singapore Convention & Exhibition Centre, Singapore
@furst_fly
Let’s connect at HUPO 2026!
Dhruv Agarwal retweeted
Self-driving labs are not about replacing scientists, they are about enabling scientists to discover dramatically more and faster.
Core Vision is human defines the scientific questions, AI decides what experiments are most informative. Robots execute them, sensors generate the data, AI learns from the results and choses what to do next.
A major focus of Prof. Milad Abolhasani’s group at NC State is the self-driving fluidic lab. These systems use continuous-flow reactors, robotics, sensors, and AI to run chemistry automatically. Because reactions occur in tightly controlled fluidic systems, they can use very small quantities of chemicals, reduce waste, and perform hundreds or potentially thousands of experiments rapidly. One of the most important innovations discussed is flow-driven data intensification. Instead of running one experiment and getting one final measurement, the system continuously monitors the reaction while conditions change. It is the difference between taking a single photograph of a reaction and recording a high-speed movie. Every “frame” becomes experimental data. This massively increases the amount of high-quality data available to AI models. Instead of generating perhaps tens of data points per day, the lab can potentially generate hundreds or thousands of measurements from a single experimental campaign. Abolhasani says this can increase useful data generation by roughly 10× to 1,000×, allowing AI models to learn chemical systems much faster and make better experimental decisions. The lab currently applies these methods to areas including quantum dots, catalysts, specialty chemicals, pharmaceuticals, energy materials, batteries, semiconductors, and quantum engineering. For quantum dots specifically, the system can optimize characteristics such as particle size, shape, composition, surface chemistry, and optical properties. The major benefits are speed, cost reduction, sustainability, and scale. Problems that historically required months or years of trial-and-error experimentation may eventually be explored in days. Automated systems can also operate continuously without breaks while consuming less material and generating less waste. However, several challenges remain. A self-driving lab is only as good as its experimental data, so data quality, noise, and measurement accuracy are critical. Another difficulty is making systems modular enough to work across very different scientific problems. Hardware and software must adapt to different chemistries, instruments, and workflows. AI decisions also need to be interpretable so scientists can understand and trust the system. Abolhasani believes self-driving labs will eventually become a standard tool in materials science and chemistry. His analogy is that within a few years, conducting advanced research without access to self-driving laboratories could feel similar to running a modern business without the internet lol.
Importantly, he does not see these systems as replacements for scientists. Instead, they are scientific co-pilots. Humans will continue defining important questions, developing hypotheses, interpreting discoveries, and understanding their broader significance. Robots and AI will handle repetitive, high-throughput, and data-intensive experimentation. That could radically increase scientific productivity. Instead of a PhD student spending four or five years solving one major experimental problem, the combination of AI and automation might allow researchers to investigate 10–20 problems during the same period.
Dhruv Agarwal retweeted
rule - before we automate anything in a wet lab @cogworksai , we go feel the manual pain first :)
so here's 4 minutes of my real manual pipetting session, squeezed into 38 seconds.
apologies to every biologist watching my technique, the bubble at the end is my magnum opus 🫠
#wetlab_basics #manual_pipetting
Dhruv Agarwal retweeted
moving liquids around using smart and bulky robots @cogworksai ;)
Dhruv Agarwal retweeted
introducing @forgeresidency
C1: 4 weeks, 12 founders, $3.3M raised by demo day
we give you:
$150k grant pool
$200k in AI credits
the biggest villa in all of goa
a hardware lab stacked with Bambu Lab X1Cs
our founders had job offers from palantir, google and cloudflare
but they dropped out, flew to SF, and raised $1M IN TWO HOURS instead
C2 x @JoinEdgeCity
30 founders, 21 days
apply below
it's kinda crazy how brand perception changes based on the amount you post
a few months ago anthropic was posting new claude code releases almost daily, and now we haven't heard from them in a while. now openai seems to be on a posting/media streak to improve branding.
as someone who works behind the scenes, there are a million reasons why this could be the case. in all cases, posting consistency on behalf of a brand is VERY hard
Dhruv Agarwal retweeted
If we really want to speed up the scientific progress then we need to rethink the way we do science in our labs.
We are building Bangalore's Lab Automators community(inspired by Bay Area Lab Automators). Open for everyone who's exploring and building in wetlab automation/ biotech/Physical AI x Bio x Robotics.
Comment here or DM. I'll send you an invite. Now it's time to have one chapter here in India 🧪+🦾

























