Open to opportunities

Herbert Bempong

@herbertbempong

Message

I build production LLM, machine learning, and data pipelines from architecture through FastAPI deployment.

Ghana

Message

What I'm looking for

I'm looking to build production AI products where I can own architecture, data pipelines, model evaluation, and deployment while aligning technical work with meaningful business outcomes.

I've built production LLM and machine learning pipelines across architecture, data ingestion, evaluation, and FastAPI deployment.

At Fhait Solutions, I architected an ingestion pipeline for 500+ unstructured documents using Gemini 2.0 Flash and Many-Shot In-Context Learning, reaching 88% extraction accuracy without expensive fine-tuning. I also built a hybrid candidate-scoring engine that reduced manual screening time by over 90%.

As Technical Lead for Intellas UK's Techpack Programme, I define system architecture, database schemas, and modular pipelines for a stealth-stage AI product. I align technical roadmaps with business strategy and investor-readiness milestones.

Alongside AI work, I've delivered a FastAPI and SQLite billboard inventory system and built data products spanning predictive maintenance, recommendation systems, environmental analysis, and freight analytics.

Experience

Work history, roles, and key accomplishments

Current

Freelance Full-Stack Developer

Current

Impact Art Ltd

Jun 2026 - Present (3 months)

Developed a full-stack billboard inventory management system using FastAPI and SQLite, serving 124 billboard listings with real-time availability filtering. Built a CI/CD pipeline with GitHub Actions that runs pytest on every push and auto-deploys to PythonAnywhere on success.

fastAPI SQLite GitHub Actions Pytest PythonAnywhere CI CD Full Stack

Current

Technical Lead, AI Product Team

Current

Intellas UK

Feb 2026 - Present (7 months)

Serve as technical lead for a stealth-stage AI product within Intellas' Techpack Programme, defining system architecture, database schema design, and modular pipeline structure. Institute engineering best practices and steer cross-functional alignment between technical roadmap and business strategy.

System Architecture Database Design Data Migration Testing Cross functional Leadership

AI/ML Engineer

Fhait Solutions

Oct 2025 - Jan 2026 (3 months)

Architected a production-grade ingestion pipeline processing 500+ unstructured documents using Gemini 2.0 Flash and Many-Shot In-Context Learning, attaining 88% extraction accuracy. Constructed a hybrid scoring engine with FastAPI, reducing manual screening time by over 90% and orchestrating a GDPR-compliant ETL workflow with automated PII anonymization.

Gemini API fastAPI ETL GDPR Anonymization Python