Machine Learning Research Engineer

Seqera
Barcelona
Workplace: OnsiteFull timeFunction: Data Science & Machine LearningSkills: ["Ownership","Problem-solving","Operational mindset","Defending results"]

Build an improvement loop for a research engine on the Seqera Platform by designing evaluation metrics, converting real deployment run data into training corpora, and running repeated post-training cycles. Post-train small open-weight models (SFT, preference tuning, adapters) to improve generating, ranking, and executing scientific ideas at lower cost and latency. Work in-office in Barcelona with a Founding Scientist and application engineers to shape the next-gen roadmap.

Loading

Loading job details...

Preparing the role view and application actions.

FursaFursa
Seqera
Seqera
9 hours ago

Machine Learning Research Engineer

✓ Verified Job

Canonical indexed version, validated from employer's careers page.

Source: Company careers pageValidated by: Fursa AI
Last checked: 8 hours agoStatus: Live

Job Summary

Build an improvement loop for a research engine on the Seqera Platform by designing evaluation metrics, converting real deployment run data into training corpora, and running repeated post-training cycles. Post-train small open-weight models (SFT, preference tuning, adapters) to improve generating, ranking, and executing scientific ideas at lower cost and latency. Work in-office in Barcelona with a Founding Scientist and application engineers to shape the next-gen roadmap.
Location: Barcelona
Workplace: Onsite
Employment Type: Full time
Job Function: Data Science & Machine Learning

Key Responsibilities

  • •Design and build an evaluation suite that defines what “better” means for an autonomous research agent.
  • •Build a pipeline to turn messy real-world run data from deployments into training-ready corpora.
  • •Post-train small open-weight models (SFT, preference tuning, adapters) to beat baseline APIs on core scientific tasks at lower cost and latency.
  • •Run repeated training cycles on evolving data.
  • •Own the next-generation roadmap with the Founding Scientist by mapping data needs to capabilities and sequencing priorities.

Pay and Benefits

Perks:EquityHealth InsuranceLife InsuranceLearning Budget

Key Requirements

  • •Post-train open-weight language models using SFT and at least one preference-tuning method, with results validated by real users.
  • •Retrain models on changing data multiple times while handling regression and forgetting (e.g., replay, data mixing, adapter strategies, eval gating).
  • •Build task-specific evaluation harnesses, define metrics, and defend results to skeptics.
  • •Fluently code in Python and use modern ML tooling such as PyTorch and transformers (plus TRL, PEFT, vLLM or equivalents).
  • •Have an open-source track record (tool contributions and/or fine-tuned models on model hubs).
Experience:Open sourceMachine learningLanguage modelsContinual learning
Skills:OwnershipProblem-solvingOperational mindsetDefending results
Tech Stack:PythonPyTorchTransformersTRLPEFTVLLMOpen-weight language modelsSFTPreference tuningAdaptersNextflowWorkflow orchestration

Company Brief

Seqera
Develops cloud-native workflow orchestration and reproducible data pipelines for bioinformatics and computational biology. Provides tooling around Nextflow to enable scalable, portable, and collaborative scientific workflows across cloud and HPC environments.
Industry: Data Infrastructure
Company Size: Small (11 to 50 employees)
Growth: Early Stage Startup
Headquarters: Barcelona, Spain
Founded: 2018
WebsiteLinkedIn