Sense
Production data, machine signals, transactions, manuals, and operational context become usable evidence.
Lewis NDAMBIRI
Industrial Engineer | AI, Optimization & Industrial Systems | Smart Manufacturing & Supply Chain | Operational Excellence
I design data products that make production systems easier to see, forecast, and improve.
100%
CNC RAG page hit rate@5 on eval set
839
Haas CNC manual chunks indexed locally
9.23%
MAPE on retail demand forecasting
31.77x
HPC speedup using MPI/OpenMP
Selected Projects
Python, FAISS, Sentence Transformers, Ollama, Mistral, Streamlit, Pytest
Offline CNC RAG Assistant
Built a local retrieval-augmented generation assistant for CNC machine operation and maintenance documentation, using a Haas Mill NGC manual as a realistic manufacturing corpus.
Result: 839 indexed manual chunks, hybrid retrieval, cited local answers, Streamlit UI, and 100% page hit rate@5 on the evaluation set.
Python, XGBoost, Pandas, Scikit-learn, Streamlit, Docker
Demand Forecasting
Built an end-to-end forecasting pipeline for the Rossmann Store Sales dataset, covering 1,115 stores and 1M+ rows with leakage-conscious feature engineering and recursive future forecasting.
Result: 9.23% MAPE, 0.914 R², and 55% RMSE reduction versus the best naive baseline.
React, TypeScript, FastAPI, MQTT, Docker, OPC UA, IIoT
Remote Real-Time Industrial Production Monitoring & Control
Built a real-time industrial monitoring platform with live telemetry, OEE analytics, alarms, audit logging, and secure remote machine control using FastAPI, MQTT, and Docker.
Result: Real-time MQTT telemetry ingestion, OEE and downtime analytics, role-based machine control, audit logging, and a containerized FastAPI–React stack.
Python, Scikit-learn, Pandas, Statistics
Maintenance Mode Prediction
Built a decision tree classifier to predict industrial machine operational states using sensor and maintenance data, with tuned hyperparameters and interpretable feature importance.
Result: 96% accuracy with key predictors including Sensor5, Age, and Sensor2.
Python, Pandas, NumPy, XGBoost, Matplotlib, Seaborn
Industrial AI Challenge - Fedrigoni
Analyzed and modeled the slitting process at Fedrigoni's Arco plant to support predictive production planning and improve manufacturing decision-making.
Result: forecasted daily slitting output and helped shift planning from reactive to predictive management.
Other Projects
Python, Flask, PostgreSQL, Docker, REST, Telegram Bot
UrbanPulse Event & Logistics Planner
Built a process-centric, four-layer service-oriented architecture that helps users discover events and receive weather-aware transport recommendations through a Telegram chatbot.
Result: integrated event, weather, and routing APIs into a containerized event planning workflow.
PostgreSQL, Python ETL, Advanced SQL, Power BI, Grafana, Docker
Retail Analytics SQL Pipeline
Built a reproducible analytics pipeline that generates synthetic retail data, loads a PostgreSQL star schema, answers business questions with advanced SQL, and publishes executive and operational dashboards.
Result: 10,000 validated transactions, cohort retention, RFM segmentation, KPI analysis, Power BI-style reporting, and a provisioned Grafana dashboard.
C, MPI, OpenMP, PBS, Linux HPC Cluster
Parallel Numerical Integration Using Romberg's Method
Parallelized a high-accuracy numerical method across MPI, OpenMP, and hybrid configurations, then benchmarked scaling behavior across cluster placements.
Result: 31.77x speedup on 32 cores with 99.3% efficiency.
React, Node.js, Express, Google Calendar API, Nodemailer, CSS
MeetScheduler AI-Assisted Meeting Scheduler
Built a Doodle-inspired meeting scheduler for HackaPrompt AI 2026, with poll creation, availability voting, quorum-based best-slot detection, Google Calendar availability, and email invitations.
Result: A full-stack scheduling prototype, demonstrating how AI-assisted development accelerates delivery when guided by strong product and engineering judgment.
Technology Stack
AI & Data
Python
Pandas
NumPy
Scikit-learn
XGBoost
FAISS
Sentence Transformers
Mistral
Ollama
Industrial & Operations
Optimization
Forecasting
Simulation
Lean Systems
Predictive Maintenance
Industry 4.0 / 5.0
Systems & Backend
PostgreSQL
Advanced SQL
SQLAlchemy
Flask
REST APIs
Node.js
Express
Docker
Docker Compose
Interfaces & Delivery
Streamlit
Power BI
Grafana
React
Google Calendar API
Telegram Bot
Nodemailer
Engineering Practice
Pytest
CI
Linux
C
MPI
OpenMP
PBS
Education
Data Science and AI for Competitive Manufacturing
EIT Manufacturing Master School, 2026. A double-degree program combining data science, artificial intelligence, industrial engineering, and manufacturing innovation.
M.Sc. Industrial Engineering
Centrale Nantes, 2026. Operations research, enterprise modeling, simulation, innovation engineering, and industrial project management.
M.Sc. Computer Science
Università degli studi di Trento, 2026. HPC for data science, AI and innovation, service design, and ICT entrepreneurship.