Computer Science · Software Engineering · ML & NLP
I'm a Computer Science graduate from Maseno University. I build backend systems and mobile apps in Go, Python, Flutter and spend time on NLP research for low-resource African languages .
Portfolio: jomboi8.davida.ink
NLP and machine learning for low-resource African languages, especially cross-lingual transfer and community-driven data curation. Most of this comes from working on Kidaw'ida (Taita), an agglutinative Bantu language that exposes where a lot of standard NLP tooling breaks down.I'm working on my interests in morphologically aware machine translation for languages like Taita, including how similarity to Kiswahili and lightweight fine-tuning methods could make translation practical for communities with limited digital resources.
TaitaCLIR — Cross-lingual semantic search for Taita. A multilingual sentence transformer (paraphrase-multilingual-MiniLM-L12-v2, later LaBSE) encodes queries and a product catalog in English, Kiswahili and Taita; FAISS retrieves the closest matches by meaning; a Streamlit interface shows results across all three languages. Built during the Princeton in Kenya Global Seminar Python · Sentence-Transformers · FAISS · Streamlit ·
Dawida — Mobile platform for building digital language resources for the Taita language. Contributors add words, meanings, example sentences and pronunciation audio, and the community verifies entries through voting, building a shared, checked lexicon Play Store . Flutter · Dart · Go · PostgreSQL · Offline-first
| Project | What it does | Stack |
|---|---|---|
| Kairo | Real-time, multi-user project & task tracker. Kanban board, calendar, Gantt-style timeline and typing indicators over a per-project WebSocket hub. Live demo. | Go, React, PostgreSQL, WebSockets |
| Mawora | Mobile-first farm management app replacing paper records with structured tracking of production, inventory, sales and expenses across multiple farm sites, offline-first with role-based access. | Flutter, Dart, Offline-first |
| EchoStrike | Single-binary Go tool that generates high-fidelity syslog traffic (RFC 3164/5424, UDP/TCP/TLS) to stress-test SIEM ingestion, parsing and detection rules, so testing security pipelines doesn't require waiting for a real attack. | Go |
Machine Learning & NLP — Sentence Embeddings, Sentence-Transformers, FAISS (Vector Search), Cross-Lingual Retrieval, Streamlit, Jupyter Notebooks
Programming Languages — Go, Python, TypeScript, SQL, Dart, Bash, C
Backend & Systems — REST APIs, WebSockets, Distributed Systems, Concurrency (Goroutines), Gin, Chi, JWT / OAuth2
Databases & Data — PostgreSQL, SQLite, Supabase, NeonDB, sqlc, Schema Design
Mobile Development — Flutter, Dart, Offline-first Sync
Cloud, DevOps & Security — Docker, Git/GitHub, CI/CD, Linux, OWASP Top 10, Burp Suite, Wireshark
BSc Computer Science, Maseno University (2022–2026).