Professional Summary
Data Scientist, AI Engineer and Data Analyst combining 10+ years of financial services experience with hands-on expertise building LLM-powered analytics platforms, RAG architectures, and production-grade automation systems. I have deployed AI solutions that reduced manual processing by 80%, built data pipelines processing UN treaty data across 180+ countries, and co-authored peer-reviewed research in predictive analytics.
My technical foundation spans Python, SQL, LangChain, scikit-learn, Flask, and AWS — applied to real problems in federal compliance, international policy, credit risk, and agricultural AI. I am particularly focused on the space where rigorous data engineering meets business accountability: systems that are not just intelligent, but trustworthy and explainable to the stakeholders who depend on them. Currently completing an MS in Analytics at American University (May 2026) and open to full-time roles in data analytics, AI engineering, and applied ML.
Academic Journey
Professional Experience
A timeline of roles across AI engineering, data science, and financial services.
Research Associate Part-time
Institute on Disability and Public Policy | Feb 2026 – Present · Washington, DC · Hybrid- Building LLM-powered analytics infrastructure for the CRPD Global Disability Rights Dashboard, analyzing UN treaty implementation across 180+ countries ahead of the June 2026 UN Conference of States Parties launch.
- Automating data collection pipelines for UN disability rights reports using Python and R ecosystems.
- Conducting NLP and transformer-based text analysis to extract policy insights from multilingual treaty documents.
- Developing interactive Shiny and Streamlit dashboards that translate complex policy data into actionable insights for civil society organizations and disability advocates.
- Contributing to academic publications; collaborating with the World Bank Disability Data Hub and Fordham's Disability Data Initiative.
Event Data Coordinator
American University | Aug 2025 – Jan 2026 · Washington, DC- Applied time-series forecasting and seasonality modeling to predict event demand and capacity, enabling data-driven resource allocation.
- Built NLP-based text mining pipelines to analyze survey feedback, extracting actionable themes through statistical testing and sentiment analysis.
- Developed segmentation models using clustering to identify high-engagement cohorts and retention drivers.
- Constructed reproducible analytics pipelines with versioned datasets feeding BI dashboards.
Software / Data Analyst – Intern
Harmonia Holdings Group LLC | Jun – Aug 2025 · Virginia, USA- Designed and integrated AI-powered validation and predictive analytics pipelines into a federal financial reporting application, improving data accuracy and reducing regulatory risk.
- Built real-time NLP-driven workflows featuring automated error detection, compliance checks, and self-generating audit trails.
- Architected and deployed a full-stack application on AWS (React, Flask/FastAPI, PostgreSQL), reducing manual processing by 80% through intelligent automation.
- Developed interactive dashboards with forecasting and variance analysis for grant management and financial oversight.
- Implemented multi-language voice input using NLP and WCAG 2.1 AA accessibility features.
Graduate Assistant – Research & Teaching
American University | Jan – Jun 2025 · Washington, DC- Supported advanced coursework in predictive modeling, machine learning, and applied data science.
- Contributed to research in statistical analysis and business intelligence.
Senior Internal Auditor – Business Operations
Uganda Development Bank | Jun 2023 – Jul 2024 · Kampala, Uganda- Built predictive analytics models using Python and SQL for loan delinquency detection and repayment risk segmentation.
- Engineered data quality frameworks and continuous monitoring systems, achieving a 95% compliance target through automated analytics.
- Designed Power BI dashboards for real-time audit tracking, driving a 78% issue closure rate.
- Developed SQL-based revenue assurance models to identify and recover lost revenue through anomaly detection.
Business Compliance Advisory Manager
Stanbic Bank (U) Ltd | Dec 2021 – May 2023 · Kampala, Uganda- Built SQL-based risk assessment and predictive models for KYC compliance, achieving a 96% compliance score.
- Led data-driven Anti-Money Laundering assessments and large transaction monitoring, applying pattern recognition and anomaly detection for 100% regulatory compliance.
Manager, Internal Audit
Equity Bank (U) Ltd | Jan 2019 – Nov 2021 · Kampala, Uganda- Automated audit processes through ML-informed exception reporting models, achieving 25% automation of daily reviews and reducing manual tasks by 60%.
- Spearheaded risk-based audit methodology adoption, improving audit efficiency by 40%.
- Delivered 30+ investigation reports with predictive insights and anomaly detection findings to executive stakeholders.
Senior Internal Auditor & Internal Auditor – Systems & Processes
Equity Bank (U) Ltd | Dec 2012 – Dec 2018 · Kampala, Uganda- Built foundational expertise in systems auditing, financial controls, and data integrity within a high-volume banking environment.
- Directed assurance services for key operational processes, increasing operational efficiency to an 85% average through analytical optimization.
Core Skills
Interests
Personal Goals and Motivation
As a Data Scientist and AI/ML Engineer, my core goal is to build intelligent systems that drive meaningful, measurable impact—whether that's accelerating scientific research, enhancing operational efficiency, or improving human-centered decision-making. I am deeply motivated by the intersection of data, machine learning, and real-world application, and I continuously seek opportunities to push the boundaries of what AI can do in service of innovation and equity.
I thrive on solving complex problems with data, and I'm particularly passionate about:
- Developing robust ML pipelines that are scalable, ethical, and production-ready.
- Harnessing NLP and LLMs to extract insights from unstructured information at scale.
- Designing AI solutions that align with privacy, fairness, and transparency principles.
What I'm Looking for in an Organization
I'm drawn to organizations that:
- Value innovation and invest in research and experimentation.
- Encourage cross-functional collaboration and knowledge sharing.
- Offer a strong culture of mentorship, continuous learning, and technical excellence.
- Are committed to responsible AI practices and real-world impact, especially in domains like healthcare, education, social equity, and sustainability.
Ultimately, I'm looking to join a mission-driven team where I can apply my technical expertise, continue to grow as a leader and problem-solver, and contribute to building systems that matter.
Featured Projects
End-to-end AI and data science solutions built in production environments.
Tools: Python, Streamlit, FAISS, sentence-transformers (all-mpnet-base-v2), Anthropic Claude, Groq (Llama 3.3 70B), Ollama (qwen3:8b), Plotly, Folium, Playwright + axe (WCAG audits), Posit Connect
The first NLP and AI-powered platform to make the full UN Convention on the Rights of Persons with Disabilities (CRPD) reporting cycle searchable, visual, and actionable — built at American University's Institute on Disability and Public Policy (IDPP) for disability rights organizations, governments, researchers, and policy advocates.
- Multi-LLM RAG Architecture: Dual-stack combining local Ollama (qwen3:8b) for lightweight insights, Groq (Llama 3.3 70B) for RAG chat and policy briefs, and Anthropic Claude for the research-and-citation pipeline — all grounded in a sentence-transformers + FAISS vector index over the full CRPD corpus.
- 17 Interactive Dashboards: Spanning Explore (Folium choropleths, reporting timelines, country profiles, document comparison, semantic search), Analyze (article coverage, co-occurrence, medical-vs-rights-based language shift), and AI-Powered (Research Assistant and Policy Brief Generator with inline citations).
- Accessibility as Subject Matter: WCAG 2.2 AA compliance is built in — the CRPD itself mandates accessible information (Articles 9 & 21), so the platform is keyboard-navigable, screen-reader tested, XSS-sanitized via bleach, and continuously audited with Playwright + axe.
- Evidence for Advocacy: Surfaces treaty compliance trends, rights-based language shifts, and cross-country comparisons across States Parties — moving DPOs, governments, and policy advocates from anecdote to defensible, quotable evidence.
Tools: Python, Streamlit, Google Gemini (LLM), Plotly, Machine Learning, Prophet/ARIMA
Food insecurity affects millions of Americans, yet the data to act on it has historically been locked behind technical complexity. This platform unlocks 15 years of county-level data (2009–2023) for policymakers, nonprofits, and researchers — making sophisticated analysis as simple as asking a question in plain English.
- Ask anything: An LLM-powered chat interface converts natural-language questions into instant visualizations and data queries, so non-technical users can explore 50+ socioeconomic variables without writing a single line of code.
- Understand the "why": SHAP explainability and anomaly detection reveal which factors — poverty, unemployment, SNAP gaps, education — drive insecurity in specific counties, enabling targeted rather than blanket intervention.
- See what's coming: Time series forecasting (Prophet + ARIMA) projects county-level rates 3–5 years ahead, answering "Which communities will cross crisis thresholds by 2028?"
- Surfaced persistent disparities: Quantified racial and geographic inequities across 15 years — including hotspots in the Mississippi Delta and Appalachia — that persist despite broad economic improvements.
Tools: React, Flask/FastAPI, AWS, PostgreSQL, NLP, Python
End-to-end AI automation platform integrating predictive analytics, NLP-driven validation, and real-time compliance monitoring for federal financial reporting. Built at Harmonia Holdings Group LLC in partnership with a cross-functional team.
- AI Validation Pipelines: Designed intelligent workflows for automated error detection, compliance checks, and self-generating audit trails that flag discrepancies before regulatory submission.
- Full-Stack on AWS: Architected and deployed the application (React frontend, Flask/FastAPI backend, PostgreSQL), reducing manual processing by 80% through intelligent automation.
- Predictive Dashboards: Developed interactive dashboards with forecasting and variance analysis for grant management and financial oversight across reporting cycles.
- Accessible by Design: Implemented multi-language voice input using NLP and WCAG 2.1 AA accessibility standards, ensuring inclusive access for all users.
© All rights reserved to Harmonia Holdings Group LLC
Tools: Python, Claude API (Anthropic), Serper API, BeautifulSoup, GitHub Actions, Gmail SMTP
Fully autonomous AI agent that searches 50+ sources daily — including 9 ATS platforms (Greenhouse, Lever, Workday, Ashby, iCIMS, SmartRecruiters), LinkedIn recruiter posts, UN/World Bank portals, and YC startups — to find Data Science, ML, AI, and Analytics roles tailored to my exact skill profile.
- AI-Scored Matching: Claude evaluates every listing against my resume — scoring skill overlap, role fit, seniority, and visa eligibility — delivering only high-confidence matches (8+/10).
- Auto-Discovery: When a search result hits an ATS the system recognizes, it automatically scrapes that company's full job board — catching roles never explicitly configured.
- Workday JSON API: Bypasses browser rendering entirely by hitting Workday's hidden API endpoint, scraping Fortune 500 career pages (NVIDIA, Intel, Nasdaq, and more) in seconds.
- Daily Email Digest: A polished HTML email arrives each morning with scored matches, skill-gap analysis, and direct ATS apply links — no aggregator noise.
Tools: Python, FastAPI, LangGraph, Claude Sonnet, GPT-4 mini, Google Calendar API, WhatsApp Business API, PostgreSQL, Groq Whisper, WAXAL (HuggingFace)
AI-powered WhatsApp assistant that automates appointment booking for service-based businesses across East Africa. Customers text or send voice notes in their local language — the system books, reschedules, cancels, and answers FAQs entirely through WhatsApp, with no app download required.
- Dual-LLM Architecture: Claude Sonnet classifies intents and extracts entities with structured output, while GPT-4 mini generates natural, friendly conversational responses.
- Multi-Language Voice Support: Supports Luganda, Runyankore, Kiswahili, Kinyarwanda, and English — using Google's WAXAL fine-tuned ASR/TTS models for African language accuracy.
- LangGraph State Machine: Directed conversation graph with persistent memory handles multi-turn booking flows, mid-conversation context switching, and client identity tracking.
- Live Calendar Integration: Real-time availability checking and event creation via Google Calendar API, with automated WhatsApp reminders 24hr and 1hr before appointments.
Certifications & Professional Training
Professional certifications and training completed across AI engineering, data science, and finance.
AI Engineer Core Track: LLM Engineering, RAG, QLoRA & Agents
This certification validates my expertise in applying advanced AI engineering techniques using Python and modern LLM frameworks. The program covered the end-to-end development of production-ready AI applications, including:
- Generative AI and LLM fundamentals (tokens, context windows, cost optimization)
- Advanced prompt engineering and system message design
- Retrieval-Augmented Generation (RAG) pipelines and vector databases
- Parameter-efficient fine-tuning custom models using QLoRA
- Building autonomous AI agents and orchestration using LangChain and CrewAI
IBM Professional Data Science Certificate
This certification validates my expertise in applying data science methodologies to real-world problems. The program covered the end-to-end development of data pipelines and predictive models, including:
- Data handling, cleaning, and exploratory data analysis (EDA)
- Relational databases and complex SQL querying
- Data visualization using Matplotlib, Seaborn, and Folium
- Building and evaluating Machine Learning models (Regression, Classification)
- Hands-on application with Python, Pandas, Scikit-learn, and IBM Watson
Financial Modeling & Valuation Analyst (FMVA)
This certification validates my expertise in advanced financial analysis, valuation, and forecasting. The program equipped me to build robust, industry-standard financial models from scratch, including:
- Structuring dynamic 3-statement models with complex, multi-year financial projections.
- Executing advanced corporate valuations using DCF, Comparable Companies, and Precedent Transactions to determine intrinsic value.
- Performing critical scenario and sensitivity analyses to assess risks and evaluate accretion/dilution in M&A.
- Analyzing financial health through in-depth credit metrics, ratio analysis, and comprehensive debt schedules.
- Synthesizing complex data into impactful PowerPoint pitch decks and interactive Excel dashboards for executive leadership.
Business Intelligence & Data Analysis (BIDA)
This certification validates my expertise in applying business intelligence tools and data modeling techniques. The program covered the end-to-end development of analytical workflows, including:
- Data transformation and automated pipelines using Power Query
- Interactive data visualization and dashboarding with Power BI and Tableau
- Database querying, aggregation, and data extraction using SQL
- Statistical foundations, predictive analysis, and Monte Carlo simulations
My Resume
Download a professionally formatted PDF copy of my resume, or view the full version online.
Let's Connect
I'm open to new opportunities, collaborations, and conversations about Data Science and Artificial Intelligence-(AI). If you think we could work together or simply want to exchange ideas, I'd be glad to hear from you.
