Noble Eselase Vulley
AI / ML Engineer & Applied-ML Researcher
Co-founder, AfricodeLab
Experience
AI / ML Engineer
2023 – PresentAfricodeLab
- —Built and deployed LLM solutions for document understanding, conversational assistants, and intelligent automation.
- —Fine-tuned transformer models on domain data and implemented Retrieval-Augmented Generation with FAISS for real-time proprietary retrieval.
- —Developed end-to-end ML pipelines: preprocessing, training, evaluation, and cloud deployment.
AI / ML Research Engineer & Project Supervisor
Sep 2025 – Dec 2025Data Intelligence and Swarm Analytics Laboratory (DISAL)
- —Led MedGemma fine-tuning for diabetic-retinopathy detection, chest X-ray classification with automated radiology-report generation, and equitable dermatological diagnosis on African skin tones.
- —Owned model-architecture decisions, dataset preparation, evaluation strategy, and ethics review across 10+ healthcare-focused ML projects.
- —Supervised intern teams delivering HR, academic, school, hostel, and admission management systems in Python, React, and PHP.
Data Analyst
2021 – 2025Independent consultancy
- —Studied the socio-economic and ecological impact of sand mining in the South Tongu District using EPA methods and inferential statistics.
- —Analyzed household energy use, LPG adoption, and the feasibility of biomass-briquette alternatives.
- —Built a fuel-price prediction model for Ghana (Jun 2023 – May 2024) using time-series analysis in Python.
Senior Software Developer
2023 – 2025AfricodeLab
- —Designed a full microservices ecosystem: Netflix Eureka service discovery, an API gateway, and centralized Spring Cloud Config backed by Git.
- —Built a centralized authentication and authorization server for secure service-to-service communication.
- —Developed core domain services including Delivery Management and Event Processing.
Associate Software Developer
2023 – PresentAfricodeLab
- —Served as lead developer for a national-scale Election Management System, from system design to implementation.
- —Built Payroll Management and Payment Voucher systems with full reporting.
- —Implemented API governance with WSO2 and traced production issues with Glowroot APM.
Database Manager & Administrator
2023 – PresentAfricodeLab
- —Designed scalable database architectures for client-facing applications on MySQL, Oracle, and PostgreSQL.
- —Cut query execution times by up to 40% through indexing, query rewriting, and schema optimization.
- —Established backup, recovery, and disaster-recovery strategies and ran live, zero-downtime migrations.
Selected Projects
Co-author / ML Researcher
Co-developed a hybrid BPSO-PSO optimized MLP for feature selection and hyperparameter search; published in Springer LNNS.
Python · Neural Networks · BPSO / PSO · Feature Selection
ML Engineer
Designed the multi-agent reasoning and hybrid RAG over policy documents, producing explainable, analyst-facing summaries.
Python · LLM agents · RAG · Qdrant · PostgreSQL · React
ML Engineer
Built the pgvector RAG pipeline and a provider-swappable LLM/embedding layer for bilingual, document-grounded answers.
Python · FastAPI · pgvector · Redis · RAG · WhatsApp Cloud API
Geospatial ML Engineer
Built the LightGBM change-detection model over Sentinel-2 imagery for sub-hectare deforestation detection.
Python · FastAPI · PostGIS · LightGBM · Sentinel-2 · Next.js
ML Researcher
Developed a graph-pooling Vision Transformer for multi-magnification histopathology with k-fold validation and ablations.
PyTorch · Vision Transformers · Graph Pooling · scikit-learn
ML Engineer
Built the RandomForest surge-forecasting pipeline with automated six-hour retraining.
Python · FastAPI · scikit-learn · WebSocket · Flutter · Redis
Publications
A Hybrid BPSO-PSO Optimized Multilayer Perceptron Framework for Enhanced Heart Disease Prediction
Nathanael Ayirebaje Azuponga, Jumah Musah, Courage Adzewoda, Noble Eselase Vulley, Justice Kwame Appati, Michael Agbo Tettey Soli.
Innovative Computing and Communications (ICICC 2025). Lecture Notes in Networks and Systems, Springer Nature Switzerland, 2026, pp. 254–264.
DOI: 10.1007/978-3-032-30005-8_21