Founding Engineer

Råvan Talani

From Uppsala to Marseille, via machine learning and a lot of TypeScript.

Founding Engineer·TypeScript·Machine Learning·Next.js·Node.js·MCP·Python·Azure·React·Marseille·Full Stack·AWS·Founding Engineer·TypeScript·Machine Learning·Next.js·Node.js·MCP·Python·Azure·React·Marseille·Full Stack·AWS·

About

I'm a Founding Engineer with an M.Sc. in Sociotechnical Systems Engineering from Uppsala University. I've spent the last two years building a Unified API from scratch at a German startup, connecting over 20 accounting and ERP systems.

My background spans both machine learning — including a production ML model for live football betting odds — and full-stack development. Most recently I've been leading AI strategy and building an MCP server to bring AI connectivity to the Unified API.

Based in Marseille, France, where I run along the Mediterranean coast, get lost in a good book, and occasionally disappear to somewhere new on the map.

Experience

2023 — Present

Founding Engineer & AI Strategy Lead

Maesn GmbH (German Startup) · Remote / Marseille

▸Built a Unified API from scratch connecting 20+ accounting and ERP systems via REST, SOAP, and GraphQL
▸Led AI strategy; independently designed and shipped an MCP server enabling AI-to-ERP natural language workflows
▸Architected the full backend stack and drove technical decisions across the product

TypeScriptNode.jsNext.jsMongoDBPrismaAzureMCPREST/SOAP/GraphQL

2023

ML Engineer — Thesis / Production

Kambi Group · Uppsala, Sweden

▸Developed a probabilistic ML model for live football stoppage-time prediction
▸Modelled outcomes as a convolution of negative binomial distributions using TensorFlow and Keras
▸Model shipped to production and actively used for live betting odds at scale

PythonTensorFlowKerasScikit-learnAWSPostgreSQL

2022

Front-end Developer Intern

Kambi Group Plc · Stockholm, Sweden

▸Independently built an internal slot-machine monitoring interface for external casino operators
▸Rebuilt the interface using modern React with hooks, improving usability and maintainability

ReactJavaScriptHTMLCSS

2018 — 2023

M.Sc. Sociotechnical Systems Engineering

Uppsala University · Uppsala, Sweden

▸Thesis: probabilistic ML model for football stoppage-time prediction, shipped to production
▸Focus on systems design, human-computer interaction, and applied machine learning

Skills

Languages

TypeScriptPython

Frontend

ReactNext.js

Backend

Node.jsMongoDB

ML / AI

TensorFlowPyTorchMCP

Cloud & DevOps

AzureAWSDocker

Projects

Unified API for Accounting & ERP

20+ accounting & ERP systems connected

Built a Unified API from the ground up that connects to 20+ accounting and ERP systems, enabling seamless financial workflows and data synchronization.

TypeScriptNode.jsRESTSOAPGraphQLMongoDBAzure

MCP Server for AI Tool Connectivity

AI-to-ERP in natural language

Independently implemented an MCP server on top of the Unified API, enabling AI assistants to interact with accounting and ERP systems through natural language.

TypeScriptMCPNode.js

Football Stoppage Time Prediction

Shipped to production at Kambi Group

Developed a probabilistic ML model for predicting stoppage time in football matches, modelling outcomes as a convolution of negative binomial distributions. Shipped to production at Kambi Group and actively used for live betting odds.

PythonTensorFlowKerasScikit-learnAWS

In Progress

French Language Learning App

Active development

A mobile app for learning French, trying to make sense of the language that doesn't make sense. Built for more structured learning than traditional apps.

ReactTypeScriptVite

Coming Soon