Herve Twahirwa
Kigali
Summary
Senior data scientist with 5+ years of experience designing and deploying production-grade machine learning systems in banking and financial services. Strong background in recommendation systems, fraud detection, LLM-powered analytics, and quantitative financial modeling. Proven ability to translate complex business problems into scalable data products that deliver measurable impact.
Overview
10
10
years of professional experience
Work History
Data Scientist
Bank of Kigali
02.2024 - Current
- Lead data scientist delivering enterprise AI platforms supporting sales, fraud prevention, and business intelligence.
- Xsell – Next-Best-Offer & Customer Intelligence Platform
- Designed and productionized Xsell, an AI platform that recommends the most relevant banking products to customers and identifies churn risk.
- Built the recommendation engine using customer behavior, transaction patterns, and similarity-based modeling to surface next-best offers in real time.
- Developed the full Xsell platform end-to-end:
- Frontend: React (agent-facing telesales interface, insights, KPIs)
- Backend: FastAPI (real-time scoring APIs, logging, authentication)
- Delivered measurable business impact within 2 months:
- 3% churn reduction
- 4% increase in digital channel adoption, especially among new-to-bank customers
- Reduced operational costs incurred by relationship officers by improving targeting and prioritization
- Enabled automated lead generation and campaign tracking, shifting sales teams from intuition-based outreach to data-driven engagement.
Risk Consultant (Quantitative Analytics)
Empyrean Solutions
01.2018 - 01.2021
- Delivered advanced financial analytics and quantitative models to large commercial banks ($50B+ in assets).
- Asset & Liability Management (ALM) and Liquidity Risk Models
- Developed Asset & Liability Management (ALM) and liquidity risk models, including deposit repricing, decay estimation, and cash-flow forecasting.
- Built deposit analytics and FTP models that improved estimation of deposit life and funding stability.
- Reduced the gap between loan funding and deposits by enabling more accurate loan-to-deposit ratios, allowing banks to:
- Fund longer-term loans
- Avoid overly conservative balance-sheet strategies
- Increase profitability through better capital utilization
- Automated regulatory and management reporting, improving scenario analysis speed and executive decision-making.
Founder & Lead Engineer
Stuvice
01.2016 - 2021
- Founded Stuvice, a student services marketplace connecting students offering and requesting services (laundry, cleaning, moving, food delivery).
- Full-Stack Development
- Led full-stack development using Flask/Django, MongoDB, MySQL, and deployed production infrastructure on AWS.
- Implemented secure data management and began development of a matching/recommendation algorithm to personalize service discovery.
- Code: https://github.com/twahih/StuviceApp.git
Education
M.S. - Data Science
Northeastern University
Boston, MA12.2023
B.S. - Applied Mathematics & Quantitative Economics
Millsaps College
Jackson, MS05.2018
Skills
- Programming: Python, SQL, R
- Machine Learning: scikit-learn, XGBoost, LightGBM, PyTorch, TensorFlow
- LLMs & NLP: LangChain, Retrieval-Augmented Generation (RAG), Prompt Engineering
- Data & Analytics: Pandas, NumPy, Spark, PostgreSQL, MongoDB, Memgraph
- MLOps & Backend: FastAPI, Docker, Kafka, MLflow, Airflow, CI/CD, Linux
- Frontend & BI: React, TypeScript, HTML/CSS, Power BI, Tableau
- Cloud: AWS (EC2, S3)
Languages
English
Full Professional
French
Full Professional
Kinyarwanda
Native or Bilingual
Timeline
Data Scientist
Bank of Kigali
02.2024 - Current
Risk Consultant (Quantitative Analytics)
Empyrean Solutions
01.2018 - 01.2021
Founder & Lead Engineer
Stuvice
01.2016 - 2021
B.S. - Applied Mathematics & Quantitative Economics
Millsaps College
M.S. - Data Science
Northeastern University