Automation Engineer

Abstract Bio; San Francisco (CA); On-site; USD 120,000 - 180,000; Full time

Job summary

Abstract Bio is building a DNA-era lab automation platform in San Francisco. The Automation Engineer will move synthesis and QC to robots, integrate instruments, and write software that plans runs and handles failures. Your work will reduce manual steps and increase reliability across R&D and production.

You will ship pragmatic automation quickly, with 3–5 years of wet-lab automation, strong Python skills, and a hands-on hardware mindset. Expect growth into multi-site automation leadership.

Qualifications

• 3–5 years automating wet-lab workflows on liquid handlers and plate robotics.
• Proficient Python; able to develop AI tools used by others.
• Able to disassemble hardware, re-teach a deck, and 3D print fixtures.
• Understand biology workflows and why dead volumes, mixing, evaporation and contamination matter.

Responsibilities

• Automate the factory floor on robots handling synthesis, assembly, QC and sequencing prep.
• Convert R&D protocols into robotic methods.
• Connect lab instruments into structured data and run scheduling agents.
• Ensure uptime, spare parts, and monitoring to avoid downtime.
• Build cheap, open hardware and custom fixtures when needed.

Skills

Automated wet-lab

AI-native

Python programming

Hardware hacking

Throughput focus

Education

Engineering degree

Tools

PyLabRobot

Opentrons

Hamilton

Tecan

Job description

TLDR

We are building the biological inputs factory that will unlock AI for biology, starting from DNA, the first building block of every biomolecule. Demand for our first DNA product outran our capacity within months of launch, and we are now scaling 10x. As our Automation Engineer, you will build and run the machines and software behind our lab: liquid handlers, plate robotics, schedulers, instrument integrations and the AI agents that tie them together. You will take protocols that a person runs by hand and make them run without that person, then keep them running when something breaks. The best candidate has 3-5 years automating wet-lab work. They are AI-native and scrappy, ship a working automation this week rather than a perfect one next quarter, and are enough of an engineer to open the machine and fix it on the spot. Your role will evolve with time; you should be excited by that and step up when needed. THIS IS NOT A 9 TO 5.

Story and mission

The Bessemer process made steel cheap and set off the Second Industrial Revolution. AI is now doing the same to biological design, and cloud labs and robots are doing it to testing. That pushes the bottleneck to the build step: the physical inputs. Every protein and every RNA starts as DNA, and DNA is still expensive. With current suppliers, the DNA to test one therapeutic antibody design costs around $100 per data point; a good antibody needs thousands of them, and a good antibody model needs millions. Abstract Bio is building the factory for those inputs, starting from DNA, the first building block.

We brought a DNA product, 4-10x cheaper than incumbents, to market within 6 months of founding the company. With a team of three and word of mouth alone, demand quickly outran our capacity, so we are now building for true scale. This means a 10x larger production platform, adding long and plasmid DNA, and starting the first R&D on RNA and proteins.

You do not get 10x more output from 10x more hands. You get it from machines that run the protocol the same way every time, software that schedules them, and agents that watch them. Building that is the job of the Automation Engineer. Our North Star is to be the biological inputs factory that gives scientists any biomolecule, on demand: the missing part of the stack that will unlock AI for biology.

We are turning DNA into a commodity, and a commodity wins on cost and reliability. Automation is how we get both at once: every step a robot runs is a step that costs less, runs at 3am and does not vary between operators. As our Automation Engineer, you make the share of the factory that runs without a human go up every month.

You are hands-on. You write the robot methods, connect the instruments, build the scheduler and debug the run that failed overnight. You work across R&D and production. When R&D proves a method at the bench, you turn it into a robot method, a schedule and a data pipeline. When production needs more throughput, you build the line that delivers it, working with the founders and the production team on what to automate next.

You are AI-native. Our lab is built so that software can read it and AI agents can act in it, and you are one of the people who make that true. You write the code that drives the instruments, connects them to each other and to our data, and lets agents plan runs, catch failures and hand off between machines.

You are scrappy. Some of what we need does not exist as a product; some of it exists but costs ten times what it should. You will hack an open platform, write the driver yourself or 3D print the fixture rather than wait for a vendor. When a robot stops mid-run, you are the one with the panel off, finding out why.

You report to the founders and work side by side with the production and R&D teams. This is a ground-floor role: as we scale and add more engineers, you will have the chance to grow into leading automation for a multi-site factory.

What you will do

• Automate the factory floor. Move synthesis, assembly, QC and sequencing prep onto robots, validated against manual runs, and raise the hands-off share every month.
• Turn R&D methods into robot methods. Work with R&D from the first prototype so new protocols are built to be automated.
• Connect the lab. Wire every instrument into structured data, then build the agents that schedule runs, catch failures and write robot methods.
• Keep the machines running. Look after uptime, spares and monitoring, so no order is lost to a robot that was down.
• Build it cheap. Prefer open hardware, your own code and a printed fixture over an expensive integration; every dollar saved comes off the cost per base-pair.

What you will need to succeed

• You have automated wet-lab work. You have 3-5 years putting lab workflows on liquid handlers and plate robotics that ran without you, and a degree in engineering, bioengineering, molecular biology or a related field, or equivalent depth.
• You are AI-native. You write solid Python, live in a terminal, and have built at least one AI tool people actually use.
• You hack hardware. You can take a robot apart, re-teach a deck, and CAD and 3D print a fixture instead of waiting for the service engineer.
• You understand the biology. You know why dead volumes, mixing, evaporation and contamination matter before you automate a step.
• You think in throughput and cost per unit. Every automation you build has a number attached.
• You move fast with composure. You ship the hack this week, stay calm when a robot crashes at 9pm, and explain trade-offs in plain language.

Nice-to-haves

• You are a trained engineer. A background in mechanical, electrical or mechatronics engineering: you can design a custom instrument, write firmware for a microcontroller, or spec a machined part for a supplier.
• You have built on open platforms. Hands-on time with PyLabRobot, Opentrons or similar, as well as commercial platforms (Hamilton, Tecan, Beckman), acoustic dispensers and plate handlers.
• You have integrated a workcell. You have connected robotic arms, schedulers, plate stores and instruments into one system and wired it into a LIMS or data pipeline.
• You have automated DNA work. Oligo synthesis, DNA assembly, PCR setup, normalisation or NGS library prep at scale, and you know where these protocols fail when a robot runs them.
• You have shipped under a quality system. Experience automating in an ISO or similar production environment, including method validation.