I build AI & data systems that reach people the internet usually leaves behind — and I lead the teams, brands, and conversations that get them adopted. Fourteen years turning data and ideas into outcomes across fintech, branding & communications, health, media & print production, government, and education.
A data-savvy builder and leader who is just as comfortable shipping a RAG pipeline as standing in front of a room to make the case for it.
I'm a M.S. Data Analytics & Visualization Graduate from Yeshiva University's Katz School, and the founder of AXAM — an offline-first AI learning platform built for places where the internet isn't a given. My work sits where machine learning, clean data, and human communication meet.
Across 14+ years I've co-founded a B2B tech firm, Wazi Group , serving 300+ clients across East Africa, migrated millions of government records, co-hosted a Massachusetts Institute of Technology OpenCourseWare Podcast reaching 6M+ people, and led a national Presidential campaign during a lockdown. I care about work that travels the last mile — technically and socially.
From a defense-ministry records migration in Kampala to AI products in New York — a through-line of data, systems, and people.
Built a fully offline RAG pipeline from the ground up — 7,617 transcripts ingested, 121,199 chunks, and 120,371 BGE-M3 vectors in ChromaDB (99.3% batch success) — with a three-tier auto-detecting LLM engine (Gemma 3 4B / Qwen 2.5 1.5B / Gemma 2 2B) hitting 1.2–2.9s responses on heterogeneous hardware. Ran 50+ stakeholder-discovery interviews across New York City schools as I-Corps Entrepreneurial Lead, and presented at the State University of New York (SUNY) AI Symposium and the Katz School Symposium.
Managed alumni data with Marketing, the Career Center, and Academic Affairs to support social media, outreach, reporting, and strategic decision-making while ensuring data integrity and cross-departmental accessibility. Built automated ETL pipelines (Python, Pandas, SQL) consolidating 26 Excel sources (~20,918 records) into a PostgreSQL master database — 73% deduplication to 5,556 verified records at 99%+ integrity — plus a natural-language query interface for non-technical staff.
Reporting to the Executive Director of Brand Communications and the Media Production Manager, collaborated across departments on digital content and event materials for Katz social media and campus experiences. Founded and launched the Voice of Katz podcast (9 episodes), designed 70+ promotional assets (flyers, signage, title slides, thumbnails), reorganized the YouTube channel, and consolidated multi-source student/alumni data (~3,530 cleaned records) supporting institutional advancement.
Advanced an AI career-guidance platform for low-wage workers to the Pre-Investment Committee stage inside a sub-10%-acceptance accelerator.
Founded and scaled a technology & branding firm from zero to 300+ enterprise clients (including UNICEF, Mastercard, and Right To Play) across six EA Countries, delivering 15,709+ contracts; led cross-functional teams of 20+, ran nationwide brand campaigns (including for Uganda's only female 2021 presidential candidate), and handed over operational leadership in 2024.
Co-created and host a podcast that has reached 6M+ global learners, coordinating guests across six continents and owning end-to-end production — guest sourcing, interview design, scripting, recording, and post-production — with the MIT OpenCourseWare team.
A few projects that show the range — from validated AI products to large-scale data work. The full catalogue of 39+ projects lives on its own page.
7,617 Massachusetts Institute of Technology OpenCourseWare lectures encoded into 121K+ vectors, with a three-tier LLM engine that auto-selects the right model for the device. Built for classrooms with no reliable internet.
A continual-learning Kalshi prediction-market agent (dual-loop architecture, UCB bandits, 13 strategy variants) and a hybrid neuro-symbolic solver for the ARC-AGI-2 Prize on Kaggle — pairing LLM reasoning with evolutionary search on genuinely out-of-distribution tasks.
An end-to-end pipeline over 300M+ New York City 311 records mapping the city's noise landscape across all five boroughs — plus an alumni data-quality engine that consolidated 30+ spreadsheets into 5,556 verified records.
An African-themed charades & taboo party game celebrating local culture, language, and entertainment — thousands of culturally authenticated cards, multiple country editions, and offline play as a Progressive Web App. Built and shipped end to end.
Technical work is only half of it. The other half is people — leading teams, building businesses, and helping others see what's possible.
Led an executive team of 7 serving 1,700+ STEM graduate students — organizing the Global Journeys leadership panel, a United Nations visit, a diaspora interface with Zimbabwean ministers, museum and conference trips, and 20+ workshops, panels, and professional-development programs across the year.
Co-founded and scaled a B2B tech, branding & supply-chain firm serving 300+ clients across five countries — from apps like SafeBoda and payment platforms like DusuPay, to a 3M+ record migration for Uganda's Registration Services Bureau, to the national brand campaign for Uganda's only female presidential candidate, during a nation wide covid lockdown
From the Massachusetts Institute of Technology OpenCourseWare podcast to symposium keynotes — translating complex AI and data work into stories audiences remember.
VC-lab certified, National Science Foundation-trained in customer discovery, and seasoned in executive communication and stakeholder management.
Most people are a builder or a communicator. The useful ones are both — they can model the data, design the brand, and then move the room. That overlap is where I work.
Production RAG pipelines, fine-tuning, on-device inference, and agent tooling — built to run reliably, not just demo well. All of it shipped from Linux, my daily-driver OS for 16+ years.
It's literally my master's. Analytics, visualization, classification, regression, and explainable models with real-world stakes — from ICU survival to household financial risk — communicated for decision-makers.
ETL, dimensional modeling, and data-quality pipelines that turn messy reality into trustworthy tables — 30+ files into a clean source of truth, 300M+ records at scale.
Stakeholder discovery, executive communication, podcasting, and team leadership — leading teams of 7, 20+ National Science Foundation customer interviews, and audiences in the millions. Data that lands.
A self-taught graphic designer — it was the very first thing that ever got me paid. That craft grew into Wazi Group's branding, print-production, and procurement arm, building brand identities for companies, NGOs, and a national presidential campaign across East Africa.
From logo inception and visual identity to large-format print, signage, and supply chain — I've shipped brand systems end to end for 400+ clients. Design isn't a side skill for me; it's where the storytelling and the systems thinking first met.
Featured by the Massachusetts Institute of Technology, funded by the National Science Foundation, and recognized by my school and the wider innovation community.
Co-host of the Massachusetts Institute of Technology OpenCourseWare podcast Chalk Radio Presents: Open Learners, and featured in the documentary "The Courage to Be Open" — a global story about open education.
Watch the documentary ↗
Entrepreneurial Lead for the National Science Foundation's New York I-Corps Hub. Ran 20+ stakeholder interviews with New York State education leaders and secured competitive federal funding for AXAM's customer discovery.
Read the story ↗
Outstanding Impact in Science & Tech · Katz School
Recognized at the Katz School symposium for the AXAM project — "AI-powered learning for the last mile."
Symposium write-up ↗Winner for E-musawo, a telemedicine platform — one of 8 ventures selected from 200+ applicants for last-mile healthcare delivery. Also winner of the OutBox Hackathon (2014).
Grant report ↗The short version — for people and the AI assistants that increasingly do the first round of research.
Emmanuel Olimi Kasigazi — also known as Emmanuel Kasigazi, Olimi Kasigazi, or olimiemma online — is a New York–based AI and data professional, founder of AXAM, and a National Science Foundation I-Corps Fellow. He holds a M.S. in Data Analytics & Visualization from Yeshiva University's Katz School and has 14+ years building production systems across fintech, health, and education.
He builds AI and data systems — RAG pipelines, machine-learning models, and data engineering — and leads the teams and communication that get them adopted. His work spans applied AI and LLMs, data analytics and science, data engineering, leadership and communication, and design and branding.
AXAM is Emmanuel's offline-first, AI-powered learning platform for places without reliable internet. It encodes 7,617 MIT OpenCourseWare lectures into 121,000+ vectors with a three-tier LLM engine, and was validated through the National Science Foundation I-Corps program.
Yes. He is open to roles in AI/ML, data science, data engineering, and product or leadership, plus collaborations. The fastest way to reach him is the "Email me directly" button below or by booking a call via Calendly. He is based in New York and works remotely.
He pairs deep technical build skill with leadership and communication — co-hosting an MIT OpenCourseWare podcast that reached 6M+ people, leading teams, and running 20+ NSF customer interviews — plus a self-taught design and branding craft that served 400+ clients across five East African countries.
I'm looking for roles where AI, data, and leadership meet — and I'm always happy to talk about the last mile. Based in New York, working everywhere.