Bioinformatics Engineer
At LatchBio, we are transforming how biological research is conducted. As laboratory automation, high-throughput assays, and machine learning converge, we are building the cloud infrastructure that enables scientists to store, visualize, and analyze data seamlessly. More than fifty biotechnology companies already rely on LatchBio to accelerate discovery—from drug development to bioproduction.
We are seeking a Bioinformatics Engineer to collaborate with customers and migrate their bioinformatics infrastructure onto Latch. You will design and implement scalable workflows, translate complex scientific problems into robust computational solutions, and help shape the future of bioinformatics tooling on our platform.
Primary Function
Create high-quality computational biology problems that capture how experts analyze real assays, make decisions, interpret data, and reason through complex workflows.
These problems will form the ground truth that enables future agentic systems to:
-
Understand common assays (RNA-seq, ATAC-seq, single-cell, spatial, etc.)
-
Execute multi-step computational workflows
-
Diagnose QC issues and interpret biological context
-
Produce statistically valid, reproducible outputs
-
Explain methodology and results with PhD-level fluency
This work requires real computational biology and engineering experience, not annotation, not surface-level summaries.
What You'll Produce:
Realistic Biological Scenarios
Full Expert Reasoning Chains
-
Step-by-step thought process an expert uses to solve the problem
-
Justification of QC decisions
-
Statistical and algorithmic assumptions
-
Alternative interpretations and failure modes
Complete Computational Solutions
-
Python or R code reflecting real-world pipelines
-
Workflow steps aligned with tools like Nextflow, Snakemake, or similar
-
Correct outputs, visualizations, and interpretations
These deliverables train models to reason, not just to compute.
Key Responsibilities
- Develop expert-level biological tasks and workflows that serve as training and evaluation data for LatchBio’s internal AI agents.
- Translate real-world computational biology reasoning into structured, agent-executable problems, including clear inputs, intermediate signals, and expected outputs.
- Encode biological ground truth and failure modes so agents can be trained to diagnose errors, make decisions, and recover from incorrect analyses.
- Produce standardized, machine-readable artifacts that integrate directly into LatchBio’s agentic workflow and evaluation infrastructure.
- Iterate on tasks using agent performance signals and QC feedback to improve robustness, generalization, and biological fidelity of LatchBio’s agents.
Qualifications
- End-to-end analysis of one or more spatial technologies: Seeker/Trekker (Slide-seq), MERFISH, DBiT-seq, Xenium, Visium, Stereo-seq, GeoMx, CosMx, or similar assays.
- Completed 3+ datasets from raw data to final insight for publications or industry studies with real-world impact.
- Working understanding of kit-specific numerical sanity checks, eg. QC thresholds, and when results deviate from expectation.
- Familiarity with computational tools used in spatial workflows (cell segmentation, cell typing, ligand–receptor analysis, etc.).
- Strong grasp of experimental design, hypothesis generation, and interpreting scientific conclusions from spatial-omics papers.
- Working knowledge of statistical inference (hypothesis testing, confidence intervals, estimators).
- Working knowledge of high-dimensional data algorithms (PCA, neighborhood graphs, UMAP).
Ideal Candidate
You are a scientifically fluent engineer who bridges experimentation and computation. You work independently, communicate clearly, and take ownership of end-to-end solutions. You thrive in fast-paced environments, enjoy debugging complex systems, and are energized by helping develop the future of AI agents in biology.
About the Team:
- We are working on huge problems at the most important intersection in history: biology and computation. We are building a team of world-class people, and we are all eager to dedicate a substantial part of our life to solving these problems.
- The undertaking has a non-trivial risk of failing. But if we succeed we will hugely accelerate scientific progress and aid the creation of therapies for cancer, solutions to global warming, and cures for aging.
- This is a hybrid role based in San Francisco.
Who You’ll Work With the Most:
You can learn more about us at any of these places!
Employment Logistics
This is an in-person 1099 Contractor role paid per-problem based in San Francisco, likely with hybrid flexibility after a ramp period. Meals are provided during in-office days. Visa sponsorship is possible.
Hiring Process
- Round 0: Apply with a resume and cover letter
- Round 1: Jordan (COS)
- Round 2: Brief take-home assignment
- Round 3: Kenny (CTO), Zhen (Bioinformatics), and Harihara (Bioinformatics)