Open to opportunities

BIAN KOUASSI

@biankouassi

Message

Senior AI Evaluator and Data Quality Analyst specializing in LLM quality, ambiguity, and evaluation rubrics.

United States

Message

What I'm looking for

I’m looking for a remote, high-autonomy role where I can lead AI evaluation and data quality work—building rubrics, auditing LLM/multimodal outputs for truthfulness and instruction adherence, and resolving ambiguous edge cases to help teams iterate faster.

I’m a detail-oriented AI Evaluator and Data Quality Analyst with 12+ years across analytical workflows, including 5+ years focused on ambiguous model outputs, ML quality workflows, and legal-technical documentation.

I apply multi-step reasoning to evaluate LLM and multimodal performance on truthfulness, logical coherence, and instruction following—while designing robust evaluation rubrics and authoring rigorous, structured justifications for complex edge cases.

In remote, highly autonomous environments, I drive data quality alignment by auditing model outputs (200+ weekly), building QA documentation pipelines, and improving team cross-annotation consistency—so engineering and data science teams can optimize faster and with higher confidence.

Experience

Work history, roles, and key accomplishments

Data Quality Analyst

Uptake Technologies

Mar 2024 - Jun 2026 (2 years 3 months)

Reviewed and audited 200+ weekly outputs from industrial ML pipelines for logical consistency, real-world factual accuracy, and instruction alignment. Designed and deployed evaluation rubrics for a QA team and authored structured rationales for edge cases and recurring failure patterns.

Instruction Following Evaluation Rubric Design

Technical Review Analyst

Relativity

Oct 2022 - Feb 2024 (1 year 4 months)

Audited high-volume legal-technical document datasets to find contradictions, reasoning flaws, and factual gaps missed by automated heuristics. Applied multi-step reasoning for classification decision logs and built an escalation pipeline that reduced downstream data rework by 18%.

Contradiction Detection Reasoning Flaw Identification Factual Gap Detection Contextual Edge Case Escalation Pipeline

AI Response Evaluator

Outlier AI

Jun 2020 - Sep 2022 (2 years 3 months)

Evaluated, scored, and comparably ranked thousands of LLM responses on reasoning depth, instruction adherence, factual grounding, and semantic coherence. Led multimodal/spatial evaluation and documented prompt-level comparative justifications with systematic failure analysis.

LLM Benchmarking Instruction Adherence Evaluation Factual Grounding And Hallucination Auditing Semantic Coherence Assessment Multimodal And Spatial Reasoning Evaluation Failure Analysis