knowledge_base_md/projects.md

Ibrahim Kabore - Projects & Portfolio

Current Projects

AIbrahim (IBot) — AI Personal Portfolio Chatbot

• Description: A production RAG-based chatbot that answers questions about Ibrahim's professional background, skills, experience, and projects. Deployed live on Hugging Face Spaces.
• Tech Stack: Python, Gradio, LangChain, ChromaDB, OpenAI, Docker
• Features: Hybrid retrieval (BM25 + vector), self-correction, streaming responses, email integration, meeting scheduling, analytics dashboard
• Role: Lead Developer & AI Engineer
• Live: huggingface.co/spaces/kimcomehome/AIbrahim

DakiKobo — AI Chatbot for Agriculture in Burkina Faso

• Description: An AI-powered agent using RAG to assist farmers with context-aware agronomic advice. Tailored for agriculture in Burkina Faso and Africa.
• Tech Stack: Python, RAG, QLoRA fine-tuning, CrewAI, Nvidia NeMo
• Features: Fine-tuning of open-source models (LLama) using QLoRA for domain-specific accuracy; multi-agent systems using CrewAI and Nvidia NeMo for complex query routing; automated crop disease detection
• Focus: AI for Social Good, Agriculture, Zero Hunger
• Link: rebrand.ly/AIChatbot4AGRI

MSSL — Moroccan Soil Spectral Libraries Platform

• Description: A production platform built with R and Shiny for predicting soil organic matter (SOM) using spectral data. Includes MLOps deployment for real-time predictions with uncertainty quantification and quality flags.
• Tech Stack: R, Shiny, Machine Learning (PLSR, RF, SVR, Cubist, QRF)
• Features: Parcel-level SOM prediction, CSV export, map views, uncertainty layers, farmer guide interface
• Live: mssl.shinyapps.io/mssl-soil-prediction/

Yield-Gap Modeling from Earth Observation & Climate Data

• Description: Independent research project using satellite imagery and climate data to model agricultural yield gaps across regions.
• Tech Stack: Python, Sentinel-2, CHIRPS, ERA5, RF/XGBoost, GEE
• Status: In progress (Oct 2025 – Present)

Past Projects

Business Website Brochure Generator

• Description: An AI agent that scans business websites and automatically generates professional marketing brochures.
• Tech Stack: LLMs, Web Scraping, Python

Bachelor Thesis: Drone vs. Satellite vs. LiDAR Comparison

• Description: Research comparing drone imagery (Timoulit region, Beni Mellal, Morocco) with satellite images and Leica scanner outputs. Involved 3D reconstruction and spatial data accuracy assessment.
• Tools: Agisoft Metashape, Pix4D, Global Mapper, Cyclone, ArcGIS

AI, ML, Geospatial, GIS & RS Learning Hub

• Description: Community learning platform sharing tutorials, datasets, and workflows for ML, geospatial AI applications including remote sensing, land cover mapping, and drone imagery analysis.
• Platform: YouTube (youtube.com/@geomatics101k)

Irrigo — Automated Irrigation System

• Description: An automated irrigation system designed to optimize water usage for agriculture.
• Focus: Precision agriculture, water conservation

Conferences & Workshops Attended

• Workshop on Earth Observation (EO) — CRSA & MIT (May 19, 2025): Oral Session on Inter-Storm Soil Moisture Changes
• GITEX AFRICA 2025 — Marrakech
• African Forum on Investment and Sustainability in Agriculture Value Chains — UM6P (April 17, 2025)
• Science Week 2025 — Benguerir, Morocco (February 17-23, 2025)
• Deep Tech Summit 2024 — Benguerir, Morocco (May 9-10, 2024)
• Conference on AI in Research: Hype or Hope — UM6P
• Workshop of the Latin American Soil Laboratory Network (LATSOLAN) — FAO, Online (2024)
• Open Street Data Mapping (OSM Mapping) — Uganda Community