Seshat AI
Request an assessmentAssessmentFR

Stéphane Kpoviessi

Data and AI engineer, founder of Seshat AI. I design the tools I deploy, and I deploy them for you.

See my full portfolio

My fieldsHover a cell to see what it covers.

Career and achievements.

Stage 1 of 4

Lyne et Frères

Data analyst, strategic support assignments

2023 to 2025, Benin and remote

Family business (agriculture and trade): my first deployment in real conditions, with non-technical teams.

  • Analysis of commercial, operational and market data with Python and Pandas, to support diversification decisions and indicator tracking.
  • Reports and interactive dashboards with Streamlit and web interfaces, to deliver the analyses to staff.
  • Analytical monitoring of agricultural activities, notably palm plantations and processing stages, to spot improvement opportunities.
  • Deployment of internal open source AI tools (OpenWebUI, open models) and support towards responsible use.

Skills unlocked

  • Python
  • Pandas
  • Streamlit
  • HTML, CSS, JavaScript
  • Dashboards
  • OpenWebUI
  • Open models
  • Support

Technical projects.

Autonomous waste collection robot

Designed for autonomous waste collection in environments such as beaches: detecting, classifying and tracking waste while telling it apart from obstacles, then planning a collection route with Dijkstra's algorithm.

  • YOLOv8
  • OpenCV
  • Dijkstra
  • Reinforcement learning

AceObject project. A reinforcement learning approach was explored for decision-making and obstacle avoidance.

Tech Watch Agent

Multi-agent monitoring platform: topic scheduling, multi-source collection, key point extraction, sourced summaries and targeted reports. LangGraph orchestration, pgvector semantic memory.

LangGraph, FastAPI, pgvector, React, Docker

See on GitHub

BI Retail

Sales and supply chain data pipeline: generation and ingestion, PySpark transformations, PostgreSQL warehouse, Airflow orchestration and Power BI dashboards.

PySpark, Airflow, PostgreSQL, Power BI, Docker

See on GitHub

Solar irradiance forecasting

Weather data preparation, time-based features and evaluation of forecasting models, with a REST API to expose the results.

Scikit-learn, Pandas, FastAPI

See on GitHub

Education and certifications.

  1. 2022

    Preparatory class MPSI

    Cycle préparatoire Marie Stella, Cotonou, Benin

    Mathematics, physics and engineering sciences.

  2. 2023

    Bachelor's degree in Electronics and Digital Systems

    JUNIA ISEN, Rabat, Morocco

  3. 2026

    Engineering degree in Big Data and Data Science

    JUNIA ISEN, Lille, France

Microsoft AI Engineer Career PathMicrosoft LearnIn progressSe préparer au métier d'analyste de donnéesMicrosoft, LinkedIn Learning2024Create ML Models in Microsoft AzureMicrosoft, Coursera2023Supervised Machine Learning: Regression and ClassificationDeepLearning.AI, Stanford Online2023Data Analysis with PythonfreeCodeCamp2023

Community involvement

JUNIA ALUMNI, Network Commission

  • Vice-president (September 2024 to August 2025): coordinating the network, mobilizing alumni and organizing professional events.
  • Communications officer (September 2023 to August 2024): communication and promotion of the network's initiatives.

Languages

  • Frenchnative language
  • EnglishB2, Linguaskill Business English
  • Yoruba, Fon, Gounnative languages

How I work.

Four steps, with an honest exit point at each: if AI is not the right answer, I tell you.

Write to me directly.

For a question, an idea or a project, the simplest is to write to me. To scope a precise need in your company, go through the assessment request.

Request an assessment