Looking for a job

Adejumo Matthew

@adejumomatthew

AI engineer in LLM evaluation and RLHF who evaluated 40,000+ responses and built multilingual evaluation sets for African languages.

Nigeria

Message

What I'm looking for

I'm seeking remote AI training, evaluation, or machine learning engineering roles.

As a Senior AI Data & Evaluation Specialist with Freelance, AI Training Platforms, I evaluated and ranked 40,000+ LLM responses, sustaining a 98.5% QA acceptance rate. I designed scoring rubrics that raised inter-annotator agreement from 78% to 92%, and red-teamed chat models across 12 harm categories.

At Fintech Startup, I built an NLP ticket classifier and fine-tuned a transformer for fraud-related text detection. Earlier, as Data Annotation Team Lead at AI Services Company, I led 25 annotators and introduced peer review that reduced rework by 35%; my work also includes evaluation sets and models for Nigerian Pidgin, Yoruba, Igbo and Hausa.

Experience

Work history, roles, and key accomplishments

Current

Senior AI Data & Evaluation Specialist

Current

Freelance, AI Training Platforms

Jan 2022 - Present (4 years 10 months)

Evaluated and ranked 40,000+ LLM responses on accuracy, helpfulness and safety, sustaining a 98.5% QA acceptance rate. Designed scoring rubrics and guidelines that lifted inter-annotator agreement from 78% to 92%.

LLM Evaluation RLHF Red Teaming Prompt Engineering Rubric Design Annotation Guidelines QA auditing Multilingual NLP Nigerian Pidgin (Fluent)Yoruba Igbo Hausa

Machine Learning Engineer

Fintech Startup

Jan 2020 - Dec 2022 (2 years 11 months)

Built an NLP ticket classifier (F1 0.91) that cut manual triage time by 45%. Fine-tuned a transformer for fraud-related text detection, reducing false positives by 27%.

Python PyTorch TensorFlow scikit learn Hugging Face Transformers LoRA Finetuning fastAPI Docker NLP Fraud Detection Drift Monitoring

Data Annotation Team Lead

AI Services Company

Jan 2018 - Dec 2020 (2 years 11 months)

Led 25 annotators on text and speech projects, delivering 500,000+ labeled items on schedule. Introduced two-stage peer review that reduced rework by 35%.

Data Annotation Team Leadership QA auditing Annotation Guidelines Project Management Speech and Text Data Processing