NCDC AI-Powered Curriculum RAG Assistant
Curriculum intelligence assistant built with LangChain and LangSmith delivering context-aware, source-grounded answers over public education and curriculum documents.
|
6+ years building production AI systems, cloud platforms, and data infrastructure across finance, healthcare, education, and the public sector.
I'm a Software Engineer with 6+ years of experience designing, scaling, and operating production AI systems, distributed cloud platforms, cross-platform mobile applications, and batch data infrastructure.
I work across machine learning, software development, data science, and data engineering — from research and model development (computer vision, NLP, RAG/LLM and agentic systems) through horizontally scalable microservice architectures, to batch and event-driven data pipelines that drive organisational decisions.
Currently pursuing an MSc in Computer Science at Makerere University while owning production systems end to end: architecture, data pipelines, model development, MLOps, deployment, and stakeholder communication.
class Developer:
def __init__(self):
self.name = "Brian Alabyekkubo"
self.role = "Software Engineer"
self.languages = [
"Python", "JavaScript",
"TypeScript", "PHP",
"C++", "Dart", "SQL"
]
self.passion = "Building systems at scale"
def get_current_focus(self):
return [
"RAG/LLM Systems",
"MLOps & DevOps",
"Data Engineering",
"Mobile Apps"
]
Curriculum intelligence assistant built with LangChain and LangSmith delivering context-aware, source-grounded answers over public education and curriculum documents.
Re-architected the AI voice/phone assistant platform as stateless microservices on Google Cloud Run with autoscaling, cutting release time by 80% and lifting uptime to 99.9%. DevOps & Frontend Developer
Algorithmic trading platform combining RAG, contextual bandits, and Hidden Markov Models for adaptive market decisions, with batch backtesting pipelines over historical data. Software Developer
AI-driven Sexual and Reproductive Health Education platform for adolescents, transitioning from text-only to multimodal interaction aligned with WHO CSE guidelines. AI Researcher
Secure ETL-backed database for nationwide member records with batch pipelines (Airflow, dbt) and event-driven streams (Kafka, RabbitMQ), improving decision speed by over 40%. Data Engineer
Hybrid YOLOv8 + GPT-2 pipeline generating diagnostic captions from microscopy images; reduced false positives by 34% and improved comprehension for 10,000+ patients.
Member-engagement and event-management platform for UNATU with Flutter, Django REST, ticketing, and offline-first behaviour across Android and iOS.
Church application improving member engagement and Bible-reading habits, built with Flutter and Firebase and released on Google Play and the App Store.
Structured digital platform for teenagers under the AiRSHE initiative, built with React Native for an engaging and accessible mobile experience.
Client mobile application built with React Native and published on Google Play. Provided ideas and support to the developer throughout delivery.
Mobile app delivering AI-generated diagnostic results and natural-language explanations from microscopy images.
GDSC Solution Challenge 2022 — Flutter app connecting blood donors with blood banks in Uganda.
Machine-learning model for decision-making optimisation using decision-focused learning techniques.
Hybrid YOLOv8 + GPT-2 pipeline for diagnostic captions; co-authored Grad-CAM explainable AI research.
Machine Learning concepts implementation combining Computer Vision with NLP.
ML system using SVM, XGBoost, and CNN across 22 clinical features for early detection.
Real estate price prediction using Linear Regression, Random Forest & Deep Learning.
AI-powered academic paper generator; owned DevOps and deployment on Google Cloud (GCE, VMs).
Library management system with complete CI/CD pipeline and comprehensive test suites.
Computer graphics projects using OpenGL and C++ for 3D rendering.
Partnering contributor to a Web platform for sustainable agriculture NGO in Uganda.
Contributed to Geuza Ministries digital platform, improving user engagement by 30% and cutting bounce rate by 15% through responsive design.
Led technology strategy and development, overseeing software architecture, team development, and delivery. Owned DevOps end to end: CI/CD pipelines, containerised and VM-based deployments, and production monitoring.
Re-architected the AI voice/phone assistant as stateless microservices on Google Cloud Run with autoscaling, cutting release time by 80% and lifting uptime to 99.9%. Built CI/CD and observability with GitHub Actions, Kubernetes, and Grafana; reduced deployment time from 45 to 6 minutes.
Designed and built Ameron Trader, an algorithmic trading platform combining RAG, contextual bandits, and Hidden Markov Models. Engineered batch backtesting and simulation pipelines over historical market data for rigorous strategy evaluation.
Researching and developing multimodal AI capabilities aligned with WHO CSE guidelines, transitioning the platform from text-only to multimodal interaction at scale for adolescent users.
Pursuing a Master of Science in Computer Science with focus on advanced machine learning, artificial intelligence, and software engineering research.
Built scalable RESTful APIs with Django and FastAPI, cutting integration time by 60%. Developed an AI-powered email productivity tool with custom NLP on Hugging Face. Implemented CI/CD achieving 99% uptime and Lighthouse scores of 95+.
Designed data architectures combining batch ETL (Python, SQL, Airflow, dbt) with event-driven streams (Kafka, RabbitMQ), improving decision speed by over 40%. Built Power BI dashboards cutting manual payroll-reconciliation effort by over 60%.
Led Agile/Scrum adoption across multiple systems, lifting team productivity by 30% and cutting project turnaround by 35%. Used JMeter and Selenium for load and functional testing across 85% of software delivered, reducing bugs by 40%.
Built AI and data-analysis solutions using PHP (Laravel), Python, and MySQL; cut system errors by 30% and improved performance by 20%. Sustained a 95% on-time delivery rate across 1.5 years.
Improved infrastructure and security to keep average ticket-resolution time under one hour, enabling uninterrupted remote learning for 500+ students. Ran 15 workshops lifting staff proficiency by 25%.
Studied Computer Science with focus on software development, machine learning, and data structures. Completed final year project on AI-powered malaria detection with natural language explanations.
Co-authored research paper on Grad-CAM applied to CNN diagnostic models for malaria microscopy images, improving interpretability by 25–30% at Makerere University.
Applied research on multimodal AI for adolescent sexual and reproductive health education, aligned with WHO CSE guidelines. Transitioning the platform from text-only to multimodal interaction.
airshe.orgResearch on applying digital mobile technologies to support tuberculosis screening, care, and follow-up in healthcare settings.
I'm open to opportunities and collaborations. Whether you have a project in mind | want to consult | teach me something | just want to say hello, feel free to reach out!
References available upon request.
Lecturer, Makerere University
ML Researcher, Makerere University
Public Health Specialist, Makerere University