Research Software Engineer I - Laboratory of Dr. Talmo Pereira

Staff Career•La Jolla, CA

About The Position

The Pereira Lab at the Salk Institute for Biological Studies is seeking a Research Software Engineer to build and maintain the software that turns computer vision models into biological discovery. The lab develops SLEAP, an open-source deep learning framework for markerless pose estimation used by research groups worldwide, and operates the research software, analysis pipelines, and computing infrastructure that carry model outputs through to published science. Research Software Engineers often come from science themselves — computationally minded researchers who found that they most enjoyed building the tools, software, and infrastructure that make discovery possible. The role sits at the interface of research and software engineering, but it is grounded in the research side: unlike a conventional software engineer who builds to fixed product specifications, a Research Software Engineer works shoulder-to-shoulder with scientists to turn their evolving needs into working tools. It is an ideal early-career role for someone who has done research and discovered that they love to code and build systems, and it can serve equally as the foundation of a long-term research software engineering career in academia or industry. The Research Software Engineer I will own the research software behind the Institute’s plant root phenotyping program: containerized, GPU-accelerated pipelines that take raw images from automated scanners, run deep learning pose-estimation models over them, extract quantitative traits, and deliver analysis-ready results to collaborating laboratories. The role spans the full path that data travels: the desktop acquisition software running on the scanners, Bloom — the web application and database where images and results live — the open-source Python packages that extract traits, and the deep learning models at the center of it. In practice this means maintaining full-stack web and desktop applications, releasing Python packages under rigorous continuous integration, running and improving model training and inference, and working directly with plant biologists to turn their experimental questions into reproducible analyses. The Research Software Engineer I works alongside the SLEAP team that develops the underlying pose-estimation framework, and contributes to those open-source tools as part of the role. Including a link to a GitHub profile (or equivalent portfolio) with examples of previous programming work is strongly recommended for full consideration. We recognize that code written for previous employers is often proprietary and cannot be shared publicly, so this is not required; however, candidates who can point to non-coursework repositories — such as personal projects or contributions to open-source software will be highly preferred.

Requirements

BS degree in computer science, engineering, quantitative science, or a related discipline is preferred for this position.
This is an early-career role; no prior professional experience is required.
Demonstrated software engineering ability — through internships, research projects, open-source contributions, or personal projects — is essential.
Full-stack software engineering ability, including proficiency in Python and at least one other language (TypeScript preferred), and experience building complete applications spanning back-end logic, front-end interfaces, and data storage.
Proficiency with the scientific Python stack (e.g., NumPy, pandas, scikit-learn), including the ability to build complete, tested, documented packages rather than one-off scripts.
Experience designing and querying relational databases (e.g., PostgreSQL), including schema changes and migrations.
Working familiarity with at least one deep learning framework (e.g., PyTorch) and a conceptual understanding of how models are trained, evaluated, and deployed.
Demonstrated interest in computer vision and in open-source software development.
Fluency with modern, AI-assisted software development, including the effective use of AI coding tools and large language models (e.g., Claude Code) to design, build, debug, and ship software rapidly.
Demonstrated ability to learn unfamiliar systems and technologies quickly and to stand up working services and pipelines through code, scripting, and automation.
Comfort working in Linux/Unix environments with version control (Git), testing, and code review.

Nice To Haves

Master’s degree or post-baccalaureate certification in a computational or scientific field.
Prior internship, research, or work experience in a scientific, laboratory, or academic environment.
Full-stack web development experience with modern frameworks (e.g., React, Svelte, Vue, FastAPI, Django).
Desktop application development, particularly with Electron or a comparable framework.
Experience integrating software with scientific instruments or hardware, such as cameras, scanners, or data-acquisition devices.
Experience with computer vision or image analysis, particularly on large scientific image datasets.
Experience with containerization (e.g., Docker) and with scheduling work on GPU resources (e.g., Kubernetes, SLURM); prior high-performance computing or systems-administration experience is welcome but not required.
Experience deploying and maintaining self-hosted applications and services, including on cloud platforms (e.g., AWS, GCP).
Experience maintaining an open-source software project, including packaging, releases, and responding to community issues.
Academic coursework, research experience, or demonstrated interest in plant science, biology, neuroscience, or another life-science domain.
A track record of shipping small tools or projects quickly and independently, and of picking up new languages, frameworks, and systems with minimal ramp-up.
Strong communication skills and the ability to work directly with scientists to translate their needs into technical solutions.
A public code portfolio (e.g., GitHub) that demonstrates initiative beyond coursework, such as personal projects or contributions to open-source software.

Responsibilities

Develop, maintain, and release open-source scientific Python packages for pose-estimation-based phenotyping, following professional software engineering practices including version control (Git), code review, testing, type checking, and continuous integration/continuous deployment (CI/CD).
Maintain and extend Bloom, the lab’s web application and scientific database for plant phenotyping data, spanning backend services, the API, the web frontend, schema design, and deployment on self-hosted infrastructure.
Maintain and release Bloom Desktop, the Electron/React application used on the plate and cylinder scanners to acquire images and upload them to Bloom, including its integration with camera and data-acquisition hardware.
Stand up, deploy, and maintain self-hosted systems and services on lab servers and cloud infrastructure, including database backup and restore, object storage, and TLS certificate management.
Build, operate, and improve containerized, GPU-accelerated analysis pipelines that carry raw imaging data through deep learning inference to quantitative trait outputs.
Train, evaluate, and deploy deep learning pose-estimation models, and maintain the datasets, configurations, and model registries that make training runs reproducible.
Connect analysis pipelines to Bloom through well-defined data contracts, provenance tracking, and write-back of results to the database.
Run analyses on experimental datasets for collaborating laboratories and deliver the results as documented, reproducible reports and figures.
Support and make use of shared high-performance computing resources, including GPU cluster scheduling and containerized job workflows (e.g., Kubernetes-based schedulers, SLURM).
Rapidly prototype, build, and iterate on internal applications, command-line tools, and utilities in response to evolving scientific needs.
Integrate systems and services through APIs, authentication, and automation to create seamless workflows for end users.
Work directly with biologists and other scientists to understand their experimental goals and translate them into technical solutions.
Prepare and maintain documentation, runbooks, and training materials that enable researchers to use the tools and pipelines independently.
Contribute to the lab’s open-source software community, including reviewing external contributions and responding to user issues.
Participate in project planning and status discussions, and coordinate with other Salk labs, Core facilities, and programs.
May assist with the onboarding or supervision of student trainees and interns.
Perform additional duties as assigned.