Bioinformatics AI Benchmarking Architect

AI Trainer Jobs; United States; Remote; USD 120,000 - 180,000; Full time

Job summary

AI Trainer Jobs is seeking a Bioinformatics AI Expert to design advanced benchmark dataset creation and AI model evaluation for autonomous tool-use in wet/dry biology.

You will craft information-reversal experiments, assemble clean data workspaces, and orchestrate multi-step tool runs using BLAST and other genomic tools, with rigorous CoT rubrics to grade performance.

The role emphasizes rigorous QA, reproducible datasets, and evaluation at scale across diverse biological data types.

Qualifications

• Experience designing benchmark tasks for AI evaluation in bioinformatics.
• Familiarity with real wet/dry lab biological data types (RNA-seq, FASTQ, BAM, mzML).
• Ability to create self-contained benchmarks with clear grading rubrics.

Responsibilities

• Create information-reversal scenario questions using anonymized biological datasets to infer identities, conditions, variants, or pathways.
• Build clean, self-contained zip workspace archives containing biological datasets.
• Design tasks requiring multi-step external tool execution (e.g., BLAST, genomic interval tools, variant annotators, phylogenetic software, differential expression pipelines).
• Write step-by-step resolution paths and robust grading rubrics with canonical answers.
• Ensure benchmark deliverables pass automated validation checks with deterministic correct answers.

Skills

Information-Reversal Problem Creation

Data Workspace Packaging

Multi-Step Tool Orchestration

Chain of Thought (CoT) & Rubric Eng.

Quality Assurance & Verification

Tools

BLAST

Genomic interval tools

Variant annotators

Phylogenetic software

Differential expression pipelines

Job description

AI Trainer Jobs is seeking a Bioinformatics AI Expert to design advanced benchmark dataset creation and AI model evaluation for autonomous tool-use in wet/dry biology.

You will craft information-reversal experiments, assemble clean data workspaces, and orchestrate multi-step tool runs using BLAST and other genomic tools, with rigorous CoT rubrics to grade performance.

The role emphasizes rigorous QA, reproducible datasets, and evaluation at scale across diverse biological data types.