O

onepot

AI-powered autonomous small-molecule synthesis

Machine Learning Researcher Intern - Machine Learning

InternshipUpdated
No salary listed
South San Francisco, CA, USA
In PersonFully in person at the South San Francisco lab.
AI & Machine LearningAI Research

About the job

Requirements
  • Strong fundamentals in machine learning and a deep understanding of modern empirical and experimental machine learning.
  • Experience developing novel model-training techniques or working with novel tasks or datasets.
  • Familiarity with PyTorch or other machine learning frameworks.
  • Knowledge of basic machine learning theory.
  • Ability to work across the entire machine learning stack, including data, training, inference, and deployment.
  • Ability to work extended hours and weekends as necessary.
  • Ability to work safely in an active chemistry laboratory and around scientific equipment. The position does not involve lab work, but some projects may require an understanding of lab workflows.
Responsibilities
  • Work with laboratory staff on data acquisition.
  • Train models of many different shapes and sizes.
  • Deploy models directly into experimental workflows.
  • Build and help maintain infrastructure and abstractions for training a diverse set of models, including data pipelines and system tasks such as parallelism strategies and quantization.
  • Train the next generation of chemistry models to support synthesis of previously inaccessible molecules.
Desired Qualifications
  • Experience in a startup, research group, competition team, or other environment involving significant ownership and limited resources.
  • Experience training large language models, particularly mid-training and post-training.
  • Experience with machine learning systems, including scaling, kernels, or parallelism.
  • Familiarity with chemistry models such as retrosynthesis or mass-spectra modeling, or with cheminformatics.
  • Experience with active learning or other techniques suited for low-data regimes.
  • Background in a deeply quantitative field, ideally with experience training machine learning models.

About the company

O

onepot

View
Website

Onepot.ai builds an AI-enabled platform for automated, multi-step synthesis of small molecules used in drug discovery and materials science. Its system combines robotic hardware with AI agents that use large language models to perform literature review, reaction planning, and experimental analysis, enabling autonomous, high-throughput experimentation. The platform also leverages graph neural networks for reaction prediction, closed-loop optimization for synthesis planning, and foundation models trained on large experimental datasets. This integrated approach differentiates Onepot.ai from competitors by offering end-to-end autonomous synthesis tightly coupled with AI-driven planning and data-driven optimization. The company’s goal is to speed up the discovery and production of new molecules while lowering the time and cost of traditional synthesis methods.

Company Size

11-50

Company Stage

Seed

Total Funding

$13M

Headquarters

South San Francisco, CA

Founded

2025

Get referred to onepot

See people who can refer or advise you

Benefits

  • Meal Benefits
  • Commuter Benefits
  • Health Insurance
  • Company Equity

Company News

RocketNews
Nov 19th, 2025
Onepot AI raises $13M to help make chemical drug creation easier

Onepot AI raises $13M to help make chemical drug creation easier. For Daniil Boiko and Andrei Tyrin, the idea for Onepot AI came from the same frustration. "The best ideas in drug discovery were often blocked not by biology, but by synthesis," Boiko told TechCrunch. Synthesis is the creation of new molecules by using chemical reactions. It's like a recipe or Lego pieces, where small pieces, ingredients, molecules, come together to form a wider puzzle picture, a food dish, a larger molecule. As one might expect, it's quite hard to create those small molecules that go on to build bigger ones. For Boiko, a Ph.D. candidate studying machine learning in chemistry at Carnegie Mellon (he received his bachelor's and master's in organic chemistry from a university in Russia), that meant realizing that drug hunters - the scientists who oversee drug discovery and development - were skipping promising ideas just because the chemical molecules to create the drugs seemed too hard to make. "The compounds never even got a chance to be tested," Boiko told TechCrunch. For Tyrin (who received his bachelor's in computer science at MIT), his time working on drug discovery computational pipelines made him realize how behind the world of drug discovery was. "The models could generate ideas in hours, but it could take months for the lab to catch up," he told TechCrunch. "We both saw that the world was throwing money into molecular design and almost ignoring the harder problem of actually making the molecules," Boiko said. But there was a geopolitical angle too, he continued, global supply chains are becoming vulnerable, and the U.S. is entering a trade war and innovative competition, once again, with China. "It was clear," Boiko said. "Small molecule synthesis needed to be rebuilt from the ground up in the United States." Boiko and Tyrin came together to create Onepot, a company that is home to the small-molecule synthesis lab POT-1. They also built Phil, an AI organic chemist, to help run experimental analysis to increase the process of compound synthesis for their early commercial partners. Those partners are biotech and pharma companies that are currently trying out their technology. On Wednesday, the company came out of stealth with $13 million of funding, including pre-seed money, and a seed round led by Fifty Years. "Currently, pharma and biotech companies either build entire teams of chemists in-house or work with contract research organizations overseas," Tryin said of the process for molecular synthesis. Human chemists can spend months of research to create even a...

TechCrunch
Nov 19th, 2025
Onepot AI raises $13M for drug synthesis

Onepot AI, founded by Daniil Boiko and Andrei Tyrin, has raised $13M to address challenges in drug synthesis. The founders identified that promising drug ideas were often abandoned due to the difficulty in synthesizing necessary molecules. Boiko, a Ph.D. candidate in machine learning in chemistry, and Tyrin, with a background in computer science, aim to bridge the gap between molecular design and synthesis, highlighting the vulnerability of global supply chains in drug discovery.

TechAmerica
Nov 19th, 2025
Onepot AI raises $13M to simplify chemical drug creation

Onepot AI raises $13M to simplify chemical drug creation. Onepot AI secures $13M to accelerate small-molecule synthesis with AI and automated labs, helping pharma companies develop drug compounds faster and more efficiently. For founders Daniil Boiko and Andrei Tyrin, Onepot AI was born out of the same frustration. "The best ideas in drug discovery were often blocked not by biology, but by synthesis," Boiko told TechCrunch. Chemical synthesis - the process of assembling small molecules through a series of reactions - is notoriously complex. You can think of it like cooking or building with Lego pieces: the small components must be created first before anything larger or useful can be made. And creating those small building-block molecules is incredibly difficult. Boiko, a Carnegie Mellon PhD candidate working on machine learning for chemistry (with a bachelor's and master's in organic chemistry from a Russian university), said he repeatedly saw drug discovery teams abandon great ideas simply because the molecules were too complex to make. "The compounds never even got a chance to be tested," he said. Tyrin experienced a parallel issue from the computational side. With a computer science degree from MIT and hands-on experience building drug discovery pipelines, he noticed a massive gap: "Models could generate ideas in hours, but the lab needed months to catch up," he explained. Both founders felt the same thing - that the industry was pouring money into designing new molecules while largely ignoring the more problematic bottleneck: actually making them. Add to that the geopolitical unpredictability of global supply chains and growing tensions with China, and the pair saw an opportunity to rebuild small-molecule synthesis from the ground up in the United States. Their answer was Onepot, powered by POT-1, a small-molecule synthesis lab, and Phil, an AI organic chemist who helps run experiments and accelerate compound development. The company is already working with early biotech and pharma partners who are trialling the technology. On Wednesday, Onepot AI officially emerged from stealth with $13 million in funding, including pre-seed capital and a seed round led by Fifty Years. Traditionally, drug companies must build entire internal chemistry teams or turn to large overseas CROs (contract research organisations). Human chemists can spend months synthesising a single compound - often costing thousands of dollars. This process involves endless trial and error: testing different structures, checking biological activity, reviewing toxicity, studying how the drug behaves in the body, and then choosing what to try next. According to Tyrin, the real bottleneck isn't testing the molecules - it's making them. Onepot offers a catalogue of molecules it can synthesise on demand. Clients browse, select the compounds they need, and Onepot produces and ships them in physical form - either as dry compounds or dissolved in vials or plates. Behind the scenes, the founders say the real innovation lies. They created a lab setup that lets AI agents study reaction "recipes," learn what works, and contribute to rapid hypothesis generation. "When running experiments, we capture every small detail," Tyrin explained - temperature, reagents, timing, and more. "Nothing is lost, which means the experiments are reproducible even a decade from now." The advantage is that Onepot's AI learns from fresh, real-world lab data rather than relying on mined literature. Boiko described their fundraising journey as "hectic." They met their lead investor through a warm introduction - what was supposed to be a brief meeting turned into a multi-hour deep dive into industrialising chemical synthesis. Other backers include Khosla Ventures, Speedinvest, OpenAI co-founder Wojciech Zaremba, and Google Chief Scientist Jeff Dean. With the new funding, Onepot will build a second lab in San Francisco to accelerate customer growth, expand the team, and advance its compound discovery engine. Competitors in the space include WuXi AppTec and Enamine. In the long run, Boiko and Tyrin believe their platform can make drug discovery at least twice as fast - and open the door to entirely new types of chemistry once considered too difficult or "weird" to attempt. "You're not just speeding up drug discovery," Boiko said. "You're expanding the design space for what drugs and materials can be. The drug we haven't discovered yet might already be out there, waiting for us to find it."