Available for research & engineering collaborations
Building intelligent systems with data, models, and intent.
I'm Melckzedek Kirya Kaisha — a Data Scientist, AI/ML Engineer, and Applied Researcher designing production-ready AI systems across NLP, generative models, forecasting, and decision support.
6+
Featured projects
5+
Research domains
10+
Models deployed
∞
Curiosity
About
A scientist's mindset, an engineer's discipline.
I design and ship machine learning systems that combine rigorous research with the craft of modern software engineering. My work spans deep learning, NLP, generative models, and forecasting — applied to problems that matter: career mobility, climate, emergency response, and intelligent decision support.
Applied AI Research
Generative AI, NLP, recommender systems, and time-series forecasting — moving ideas from notebooks to production.
Production Engineering
Shipping end-to-end AI products with TypeScript, Next.js, and Supabase — full-stack, scalable, observable.
Decision Intelligence
Designing models that don't just predict — they inform decisions in healthcare, climate, education, and emergencies.
Selected work
Projects at the intersection of research and product.
Tech stack
Tools I reach for, sharpened by use.
AI & Data Science
PythonTensorFlowPyTorchScikit-learnPandasNumPyJupyterStreamlit
Development
TypeScriptNext.jsReactNode.jsSupabaseFirebaseTailwind CSS
Research & Tools
LangChainHugging FaceOpenAI APIGeminiDockerGit
Research interests
Questions I keep returning to.
01
Generative AI
Diffusion, GANs, and LLM-driven synthesis for low-resource contexts.
02
AI for Sustainability
Climate, weather, and resource forecasting in African geographies.
03
Decision Support Systems
Models that augment human judgment in healthcare, education, emergencies.
04
NLP & Recommendation
Multilingual NLP and contextual recommender systems.
05
Time-Series Forecasting
Sequence models, ConvLSTMs, and probabilistic forecasting.
06
AI Automation
Agentic workflows and tool-using LLMs for real-world automation.
Contact
Let's build something intelligent.
Open to research collaborations, freelance AI engineering, and full-time roles.