Open to opportunities

Japheth Kimeu

@japhethkimeu

AI Evaluation Specialist and CS PhD candidate optimizing LLMs—improving accuracy, safety, and prompt performance through rigorous evaluation.

Canada

Message

What I'm looking for

I’m open to remote roles (40+ hours/week) where I can evaluate and improve LLMs, build safer prompt workflows, and produce structured metrics and actionable reporting while maintaining confidentiality and fast, reliable delivery.

I’m an AI Evaluation Specialist and Computer Science PhD candidate with 4+ years of experience optimizing large language models (LLMs), engineering high-performing prompts, and delivering precise data annotation. I’ve published research in medicalAI (Elsevier’s Informatics in Medicine Unlocked) and focus on making models more reliable, safer, and more accurate.

I evaluate and rate 500+ AI-generated responses weekly, design and refine 100+ prompts, and use structured metrics to guide iteration. Through iterative feedback frameworks, I’ve improved model accuracy by over 25% while reducing hallucination rates by ~25% and identifying biases, logical inconsistencies, and safety violations.

I also combine technical depth with disciplined execution—handling high-volume remote workflows, enforcing strict confidentiality, and maintaining a 100% on-time submission rate. I’m proficient in Python, PyTorch, TensorFlow, and major annotation/QE tools, and I’m committed to producing clear, actionable reporting that helps engineering teams ship better systems.

Experience

Work history, roles, and key accomplishments

Current

AI Evaluation Specialist

Current

Outlier, Remotasks, RWS

Jan 2022 - Present (4 years 5 months)

Evaluated and rated 500+ AI-generated responses weekly for safety, factual accuracy, and human preference alignment. Designed and refined 100+ prompts, reducing hallucination rates by ~25% through iterative feedback, while maintaining 98%+ accuracy on high-volume data annotation and labeling.

LLM Evaluation Prompt Engineering Data Annotation AI Safety Python PyTorch Quality Assurance

Graduate Intern

West Meru Hospital

Mar 2023 - Aug 2023 (5 months)

Analyzed chest X-ray metadata with Python (NumPy/SciPy) to identify improvement areas, increasing project feasibility by 90%. Trained staff on LungGuard and created documentation/tutorials, reducing onboarding time by 30%, and contributed to a deep learning pneumonia detection solution published in a peer-reviewed journal (Elsevier).

Medical Imaging Chest X Ray Analysis Python NumPy Scipy Deep Learning Technical Documentation Staff Training

Weighbridge & Security Officer

LadyAskari

Mar 2021 - Jan 2022 (10 months)

Conducted physical verification counts and random checks, coordinating with security and logistics teams to reduce unauthorized access incidents by 20%. Produced daily weighbridge reports and implemented secure data tracking and operational logs to support compliance and audit requirements.

Security Coordination Operational Reporting Data Tracking Audit Support Document Generation

Teacher and Facilitator

Darajani Secondary School

May 2018 - Dec 2018 (7 months)

Revised curriculum delivery and lesson approaches to improve student understanding and performance by 50%. Implemented debate-style learning and individualized support plans, boosting engagement by 70% and raising struggling students’ performance by 70%.

Lesson planning Student Mentoring Debate Based Learning

IT Attache

Kenya Civil Aviation Authority

May 2018 - Aug 2018 (3 months)

Analyzed organizational workflows with SPSS and documented findings to help senior staff plan and prioritize work. Built and deployed Java-based workflow automation and an asset management system using SQL, improving asset tracking by 90% and increasing service delivery by 50%.

Workflow Analysis SPSS Java SQL Automation Asset Management Technical Writing Process Improvement