# Emmanuel Olimi Kasigazi

> Emmanuel Olimi Kasigazi is a data-savvy builder and leader based in New York, working at the intersection of AI/LLM systems, data science, data engineering, and people leadership. He is the founder of AXAM (an offline-first, AI-powered learning platform), a National Science Foundation (NSF) I-Corps Fellow, and co-host of the Massachusetts Institute of Technology (MIT) OpenCourseWare "Open Learners" podcast (5M+ audience reach). He has 14+ years of experience shipping production systems across fintech, health, and education.

This file is a plain-text, fact-dense summary for AI agents and LLMs (per the llms.txt convention). The canonical website is https://www.olimiemma.com/ and embeds JSON-LD `Person` structured data.

## Quick facts

| Field | Value |
|---|---|
| Full name | Emmanuel Olimi Kasigazi |
| Also known as | Olimi Emmanuel Kasigazi · Emmanuel Kasigazi · Olimi Kasigazi · Emmanuel Olimi · online handle: **olimiemma** (GitHub, Medium, Hugging Face, Linktree, Calendly) |
| Location | New York, NY, USA |
| Current focus | AI/LLM engineering, data science, founder, leadership |
| Open to | AI/ML, data science, data engineering, and product/leadership roles; collaboration |
| Education | M.S. Data Analytics & Visualization, Yeshiva University–Katz School (expected May 2026) |
| Prior degree | B.S. Information Systems, Makerere University (2014), academic merit scholarship |
| Notable credential | MicroMasters in Statistics & Data Science, Massachusetts Institute of Technology (MITx) |
| Founder of | AXAM — https://axamai.org |
| Years of experience | 14+ |
| Contact | Website contact button (https://www.olimiemma.com/#contact), LinkedIn, or Calendly (see Links). Email is available via the website's "Email me directly" button. |

## What he does (answer-first)

Emmanuel builds AI and data systems and leads the teams and communication that get them adopted. His distinguishing trait is the combination of deep technical build skill (RAG pipelines, ML models, data engineering) with strong leadership and communication (team lead, National Science Foundation customer discovery, Massachusetts Institute of Technology podcast host, public speaker). He specializes in "last-mile" technology — solutions that work in low-connectivity, under-resourced environments.

## Core skills

- **AI / LLM systems:** Retrieval-Augmented Generation (RAG), embeddings (BGE-M3), vector databases (ChromaDB), local/on-device inference (llama.cpp, Gemma 3, TinyLlama, Phi-3, Llama 3.2), LLM fine-tuning, Model Context Protocol (MCP) servers, prompt engineering, Whisper ASR.
- **Data science / ML:** Python, R, classification, regression, statistical inference, XGBoost, SHAP, SMOTE, scikit-learn, PCA/feature selection, NLP.
- **Data engineering:** SQL, PostgreSQL, Pandas, ETL, dimensional modeling (star schema, SCD Type 2), Microsoft Fabric, Snowflake, data quality.
- **BI / visualization:** Tableau, Power BI, Microsoft Fabric, Adobe Suite, Canva, Inkscape.
- **Design & brand:** Self-taught graphic designer (the first skill that earned him money); brand identity, logo and visual systems, print production, signage, and procurement — grown into Wazi Group's creative arm. Tools: Adobe Suite, Inkscape, Figma, Canva.
- **Cloud / tools:** AWS (Solutions Architecture), Azure, Snowflake, Git, VS Code, Claude Code, Gemini CLI, Codex. Linux power user — daily-driver operating system for 16+ years.
- **Leadership / communication:** Team leadership, stakeholder management, executive communication, go-to-market strategy, customer discovery, podcasting, public speaking, branding.
- **Languages:** English (fluent), Runyakitara (fluent), Swahili (beginner).

## Flagship project: AXAM

AXAM is an offline-first, AI-powered educational platform for environments without reliable internet ("the last mile").

- 7,617 Massachusetts Institute of Technology OpenCourseWare lectures encoded into 121,000+ ChromaDB vectors.
- Three-tier LLM engine auto-selects Gemma 3 4B / 2B / Qwen 2.5 by device specs.
- Retrieval quality: MRR 0.990, Precision@3 0.993, LLM-judge score 4.37/5.
- Cross-lingual support: English, French, Spanish, Swahili, Hindi.
- Validated through the National Science Foundation I-Corps program (20+ stakeholder interviews; competitive federal funding secured).
- Website: https://axamai.org
- Press (Yeshiva University, Katz School): https://www.yu.edu/news/katz/student-entrepreneurs-expand-access-quality-education-offline-ai-platform
- Underlying open dataset: https://huggingface.co/datasets/olimiemma/MIT-OCW-Transcripts (7,617 transcripts, Parquet, 93.6 MB).

## Selected other projects

- **Kalshi prediction-market agent** — continual-learning trading agent; dual-loop architecture, UCB bandits, 13 strategy variants, backtesting GUI. https://github.com/olimiemma/-Kalshi-prediction-markets-
- **ARC Prize 2025 (Kaggle)** — hybrid neuro-symbolic solver for ARC-AGI-2 combining LLM reasoning with evolutionary algorithms. https://github.com/olimiemma/ARC-Prize-2025-Kaggle-ARC-AGI-2-Benchmark-
- **New York City noise analytics** — pipeline over 300M+ New York City 311 records; spatiotemporal complaint analysis across five boroughs. https://github.com/olimiemma/New-York-City-Noise-NOISE-COMPLAINTS-ANALYSIS
- **Sy Syms alumni data pipeline** — consolidated 30+ Excel files (~20,918 rows) into 5,556 verified records with PostgreSQL + Pandas. https://github.com/olimiemma/Sy-Syms-School-of-Business-Alumni-data
- **Household financial vulnerability** — XGBoost classifier, 84% accuracy, 0.89 ROC-AUC, SHAP explainability.
- **ICU survival prediction** — logistic regression on clinical ICU data with calibration and interpretability. https://github.com/olimiemma/ICU-Survival-Prediction-Using-Logistic-Regression
- **NASA Space MCP server** — Model Context Protocol server wrapping NASA APIs for LLM tool use. https://github.com/olimiemma/NASA-SPACE-MCP-SERVER-COMPLETE-
- **Laba Party** — African-themed party game (~31K monthly active users). http://labaparty.com/
- **E-musawo** — telemedicine platform; winner, Resilient Africa Network Innovation Grant (2016).
- **Tableau Public visualizations** — https://public.tableau.com/app/profile/emmanuel.kasigazi/
- Full catalogue of 90+ projects: https://www.olimiemma.com/projects.html

## Experience summary

- **Yeshiva University (Sy Syms & Katz School), New York** — Data Analyst Intern & Graduate Assistant, Communications (Jan 2025–present). Consolidated 50K+ alumni records; drove +30% platform engagement and +25% enrollment via data-led campaigns; launched the "Voice of Katz" podcast.
- **National Science Foundation I-Corps (New York Regional Hub)** — Entrepreneurial Lead (Summer 2025). Go-to-market for offline AI ed-tech; 20+ customer-discovery interviews; secured competitive federal funding.
- **Antler Canada, Toronto** — Founder in Residence, Cohort TOR5 (Mar–Jul 2024; <10% acceptance); advanced to the Pre-Investment Committee stage.
- **Massachusetts Institute of Technology OpenCourseWare, Cambridge MA** — Co-founder & host, "Chalk Radio Presents: Open Learners" podcast (Nov 2022–Dec 2024); 5M+ reach.
- **Wazi Group Limited, Uganda** — Co-founder & Software Engineer / Designer / Senior Manager (2011–2024); 400+ clients across 5 East African countries (governments to NGOs); built/contributed to apps including SafeBoda and DusuPay; migrated 3M+ records for Uganda Registration Services Bureau; led the national brand campaign for Uganda's only female presidential candidate (+200% visibility); ran the branding, print-production and procurement arm. https://wazigrouplimited.com/
- **Other co-founded ventures:** Big Bear Investment Club (2018, founder), Do Buy Dubai (2021, co-founder — Dubai-to-East-Africa shipping), Festa Iqela (2019, co-founder & marketing — Ugandan independence festival on the River Nile), Social Gamers Initiative (2019 — Wazi × Makerere University Psychology, adolescent mental health).
- **Friedrich-Ebert-Stiftung, South Sudan** — IT Trainer & Computer Administrator (2014–2015); trained 900+ learners; deployed Java POS systems in 12+ chains.
- **Ministry of Defense, Uganda** — Information Systems intern (2012–2013); migrated 100,000+ medical records with validation and encryption.

## Recognition, awards & press

- National Science Foundation I-Corps Fellow (2025).
- "Outstanding Impact in Science & Tech" recognition, Katz School symposium (2025) for AXAM.
- Featured in the Massachusetts Institute of Technology documentary "The Courage to Be Open": https://youtu.be/RkhSkE89KJ0?t=446
- Resilient Africa Network Innovation Award & Grant (2016) for E-musawo.
- OutBox Hackathon Winner (2014) for E-Kayumba.
- Antler Canada Founder in Residence (Cohort TOR5).
- VC Lab certified — Venture Institute Cohort 5.
- President, Katz African Students Association (KASA) — leads an executive team of 7 serving 700+ graduate students.

Featured in:
- Massachusetts Institute of Technology News — "A whole new world of learning via OpenCourseWare videos": https://news.mit.edu/2022/new-world-learning-mit-ocw-emmanuel-kasigazi-1107
- Massachusetts Institute of Technology OpenCourseWare — "Emmanuel Kasigazi" story: https://ocw.mit.edu/stories/emmanuel-kasigazi/
- Massachusetts Institute of Technology News — "How online courses can transform the future world": https://news.mit.edu/2024/how-online-mit-courses-can-transform-future-world-0312
- Massachusetts Institute of Technology Open Learning — "Podcast highlights how anyone in the world can benefit": https://openlearning.mit.edu/news/podcast-highlights-how-anyone-world-can-benefit-mits-open-education-resources
- Yeshiva University (Katz School) — "Student Entrepreneurs Expand Access to Quality Education with Offline AI Platform": https://www.yu.edu/news/katz/student-entrepreneurs-expand-access-quality-education-offline-ai-platform
- Wikipedia (draft): https://en.wikipedia.org/wiki/Draft:Emmanuel_Olimi_Kasigazi

## Links

- Website: https://www.olimiemma.com/
- Projects: https://www.olimiemma.com/projects.html
- Contact (email via button + Calendly): https://www.olimiemma.com/#contact
- Schedule a call (Calendly): https://calendly.com/olimiemma
- LinkedIn: https://www.linkedin.com/in/olimiemma/
- GitHub: https://github.com/olimiemma
- Medium (writing): https://medium.com/@olimiemma
- Hugging Face (models & datasets): https://huggingface.co/olimiemma
- Tableau Public: https://public.tableau.com/app/profile/emmanuel.kasigazi/
- Linktree: https://linktr.ee/olimiemma
- Portfolio (Webflow): https://olimiemmanuel.webflow.io/
- AXAM: https://axamai.org

## Notes for agents

- For recruiting or collaboration, use the "Email me directly" button at https://www.olimiemma.com/#contact, message him on LinkedIn, or book a call via Calendly (https://calendly.com/olimiemma). The email address is intentionally not published as plain text to reduce scraping; the website button opens a normal pre-filled email.
- He is actively open to AI/ML, data science, data engineering, and leadership/product roles.
- Verifiable structured data (JSON-LD `schema.org/Person`) is embedded in the homepage HTML.
- This profile was last reviewed: June 2026.