JG Science enabler / tools / backend

Backend, data, and simple tools for scientific teams

Science enabler.

I simplify research work into small webapps, Python services, data workflows, and source-grounded AI tools that remove process drag without turning every task into a platform.

Download CV
20,000scientific articles screened in 3 hours with source-grounded review tooling.
Weeks to minutesspreadsheet and handoff bottlenecks replaced with simple internal tools.
Platformreproducible compute foundations across Docker, Ansible, Nix, Slurm, and CI.
Audit awareRBAC, ReBAC, GDPR-sensitive contexts, NIS2 and ISO 27001 readiness.

Selected work

Selected work.

Grounded in recent CV work: applied AI, bioinformatics workflows, regulated platform engineering, and reproducible scientific compute.

Evaxion / scientific tooling

Research workflows made faster and safer

  • Accelerated expert literature screening: 20,000 articles in 3 hours, compared with roughly 800 per week for a five-person manual workflow LLM-assisted screening, source-grounded RAG, LlamaIndex, Pinecone, PostgreSQL/pgvector.
  • Turned fragile research handoffs into runnable internal tools for lab data, compute, service intake, and operational visibility Streamlit, structured validation, QC plots, Ansible, Slurm, Azure AD, GitLab Service Desk, Grafana/logging.
SeeQ Diagnostics / bioinformatics

Research pipeline to operable workflow

  • Modernized an expert-operated Snakemake pipeline into an installable Python/Prefect workflow product Python packaging, Prefect, Snakemake, one-command execution.
  • Made the workflow safer to change with deterministic validation data, CI regression runs, configuration validation, and logging Validation fixtures, CI regression execution, configuration checks, structured logs.
Compliance / platform engineering

Typed platform foundations for regulated work

  • Built typed platform foundations for compliance-sensitive product work across APIs, UI, jobs, payments, and schema migrations FastAPI, React, SQLAlchemy/Alembic, background workers, scheduled jobs, payment integration.
  • Kept privileged operations server-side and modeled access across users, organizations, and shared resources RBAC/ReBAC, organization boundaries, shared-resource authorization, regulated-system constraints.
Scientific compute / reproducibility

Tools that researchers can actually run

  • Reduced lab workflow time from weeks to minutes by turning messy inputs into structured review and analysis paths Streamlit, structured input, QC plots, validation boundaries, downstream analysis.
  • Made scientific tools reproducible and runnable across local, containerized, and cluster environments uv, Nix, OCI images, Docker, Ansible, Slurm, CI, observability.

Availability

Primarily looking for full-time software developer roles.

Best fit: biotech, medtech, bioinformatics, scientific computing, internal automation, backend platforms, and regulated systems where simple tools can unlock scientific work. Focused consulting is possible when the problem is specific and technical.