Software engineer - Pipelines

Overview

Location: On-site, Cambridge, MA (strongly preferred; hybrid flexibility possible for exceptional candidates)

Level: Software Engineer or Senior Software Engineer

About the Role

We are a small, mission-driven research organisation building large scale, time critical data pipelines. Our team processes very large volumes of high throughput scientific data every week, and speed matters. The difference between a pipeline that takes hours and one that takes minutes has real consequences for the outcomes we exist to improve.

We are hiring a Software Engineer to own the performance, reliability, and observability of these pipelines. You will work alongside a small team of engineers and scientists and report to the engineering lead for the pipelines group. The team will be four people once this hire and one other open role are filled, so your work will have immediate, visible impact.

Prior experience in life sciences or bioinformatics is not required. The problems here are software engineering problems first. You will be embedded with subject matter experts and we need you to bring the engineering depth.

Pipeline performance and accuracy

• Benchmark pipelines end to end, work out where time is actually going (CPU cache locality, network I/O, algorithmic complexity), and implement optimisations with measurable impact
• Design and refactor algorithms, including full redesigns where the data supports it, not just micro optimisations
• Improve detection accuracy, balancing sensitivity and specificity in collaboration with the science team

Infrastructure, observability, and operational reliability

• Own pipeline observability: build dashboards, implement automated alerting, and expand test coverage so failures are caught early and diagnosed fast
• Manage cloud infrastructure (currently AWS, using Nextflow, Terraform, and related tooling), optimising for cost, throughput, and graceful degradation under load
• Improve validation and cleaning of incoming sample metadata, and automate more of what is manual today
• Support secure, compliant public data releases, including scrubbing of confidential and sensitive data

• 2+ years of professional software engineering experience or an equivalent background. Seniority is flexible based on demonstrated skill
• Strong performance optimisation skills across the full stack: memory access patterns, network I/O, and algorithmic complexity, with good judgement about which one is the real bottleneck in a given situation
• Experience with distributed or cloud based data pipelines at real scale, including their failure modes
• Strong programming skills in a language of your choice, with a genuine mental model of how it executes under the hood
• Creativity in algorithm design, not just implementation efficiency
• Thoughtful, current use of AI tools. You use them to move faster and do better work, but you know where they help and where they introduce risk
• Excellent written and verbal communication, including the ability to explain a system clearly to someone who is not an engineer
• Genuine excitement about the mission. This is a small team doing work that matters, and the people who thrive here care about why

• Experience with workflow orchestrators (Nextflow, Flyte, Airflow, or similar)
• Infrastructure as code experience (Terraform or equivalent)
• Familiarity with Python, Rust, or AWS
• Experience processing large scale scientific or sequencing data
• Experience designing systems for both robustness and flexibility

• You own things. You see what needs doing and start doing it, without waiting to be assigned
• You bring ideas, not just execution. We want someone who engages with the underlying problem and has opinions about how the system should evolve, not someone who only closes tickets
• You are collaborative. You will bring an engineering perspective to a team of scientists who code. That is valuable, but changes to shared systems get worked through together, not shipped unilaterally
• You are comfortable with ambiguity. This role needs someone who reasons from first principles rather than waiting for a fixed specs.