BackEmmanuel Thiéry

With a background in neuroscience and physical chemistry, I specialize in machine and deep learning using Python. I manage end-to-end AI pipelines — from data acquisition to deployment — building containerized APIs, LangChain agents, and CI/CD pipelines on AWS.

Work Experience

Chatbot SaaS Product

Oct 2025 – Current
Prophetic· Paris
  • Building Prophetic Orchestra 7.5 — a luxury investment advisor chatbot SaaS
  • Frontend: Next.js 15 App Router, TypeScript, Tailwind CSS
  • Backend: Python, FastAPI, LangChain / LangGraph, Supabase, multi-scraper with Scrapy — fully decoupled architecture
  • Observability with LangFuse; predictive model of artist valuation
PythonFastAPILangChainLangGraphNext.jsSupabaseLangFuseWeb Scraping

Process Automatisation

Dec 2025 – Jan 2026
ZénithIA· Paris
  • Developed n8n workflows to automate processing of SharePoint data, Excel files and regulatory information
  • Automated extraction and transformation of Excel data for compliance and regulation mapping (ETL)
  • Automated transcript generation from processed data for documentation and reporting
n8nETLAutomation

Chatbot Developer

Aug 2025 – Jan 2026
OMD· Paris
  • Developed FastAPI backend services: scraper, PostgreSQL retrieval, and web search tool integrations
  • Built and maintained a ChatGPT‑style frontend interface using Next.js
  • Implemented Google authentication for secure user access and session management
  • Set up LLM interaction tracking and answer evaluation mechanisms
FastAPIPostgreSQLNext.jsLLM / RAG

N8N Workflows Creation

Aug – Sep 2025
Eurus· Tokyo
  • Automated mail processing pipeline: attachment extraction, OCR, database ingestion and draft creation
n8nAutomation

Scraping & API — Accounting Directories

Jun 2025
Pont9
  • Monthly scraping of accounting firm directories
  • Data processing and ingestion into MongoDB
  • Built a FastAPI service to expose the collected data
  • Deployed on GCP (Cloud Run / Cloud Functions, Google Cloud Storage)
PythonFastAPIMongoDBWeb ScrapingGCP

POC Trading Bot — Reinforcement Learning

Mar – May 2025
Independent· Tokyo
  • Built trading gymnasium environment, dataset generation and cleaning, feature engineering
  • Trained using PPO and RecurrentPPO (stablebaselines3) with full experiment tracking
PythonReinforcement LearningData Science

Agent Chatbot Development

Dec 2024 – Mar 2025
OMD· Tokyo
  • Built a LangChain agent for art-market insight, querying a Supabase database and generating charts on demand
  • Implemented graph-generation tools (Plotly, Seaborn) wired into the agent
  • Delivered the chatbot through a Streamlit interface
LangChainLangGraphStreamlitSupabase

Reinforcement Learning Debugging

Nov 2024
Kuan-Yu Chiu
  • Debugged multiple reinforcement learning algorithms in a game theory context
Reinforcement LearningPython

AI Chatbot Development

Feb 2023 – Oct 2024
Re-mind
  • OpenAI API + LangChain framework, Streamlit prompt management app, MongoDB conversation storage
OpenAI APILangChainStreamlitMongoDBLLM / RAG

Webapp to Analyse Horse Performance

Oct 2022 – Feb 2024
EquiData Sport
  • Built a Streamlit web app letting non-technical users analyse horse performance (filter by age/horseman, Excel upload, result export, dynamic charts, auth)
  • Scraped multiple websites with Selenium on CRON-triggered AWS Lambda; cleaned, normalised and stored CSVs on S3
  • Serverless processing pipeline on AWS Lambda, GitLab CI/CD
StreamlitAWS LambdaWeb ScrapingGitLab CI/CDDashboard

Improve Document Selection System

Mar – May 2024
Legaware
  • Audited a chatbot answering questions over legal documents: code review and RAG/chatbot evaluation
  • Built evaluation datasets to measure the impact of embeddings, rerankers and chunk size
  • Stack: Qdrant vector database, OpenAI ada embeddings, LangChain, LangSmith for evaluation
LLM / RAGLangChainLangSmithRefactoring

E-Commerce AI Solutions

Apr – May 2024
OCITO LA GALERIE
  • Wrongly-categorised e-commerce products: built few-shot prompts to categorise products with an LLM
  • GPT-4-Turbo product description generation, web scraping for classification
OpenAI APIWeb ScrapingLLM / RAG

Art Market Analysis

Jan 2024 – Aug 2025
OMD
  • Scraped several marketplaces (Scrapy, Crawlab), cleaned and normalised ~1M records, designed the SQL schema and statistics
  • Exposed artist popularity/revenue rankings through a FastAPI endpoint, stored in Supabase
ScrapyFastAPIWeb ScrapingSupabaseStatistics

Chatbot + RAG Pipeline

May 2023 – May 2024
Kiklos
  • Built a chatbot answering any question about electronic devices (how to repair, specifications) from a PDF corpus
  • Created the RAG pipeline with evaluation: embedding choice, chunk size and reranker selection (RAGAS)
  • Built evaluation datasets to validate chatbot changes (LangSmith), served via FastAPI
LangChainLLM / RAGFastAPILangSmith

Stable Diffusion Deployment on Replicate

Aug – Oct 2023
MANUSCRY
  • Deployed Stable Diffusion workflows on Replicate (SD 1.5 and SDXL)
  • Evaluated several models and ControlNets across txt2img and img2img workflows
  • Built a CI/CD pipeline (GitHub Actions) to deploy these workflows to Replicate
Stable DiffusionReplicateGitHub Actions

Python Migration 3.6 → 3.8

Jan – Mar 2023
BearingPoint
  • Migrated an internal HR banking application from Python 3.6 to 3.8 (plus dependency upgrades) using CMake
PythonCMakeMigration

NLP, OCR & Cloud Specialist

Sep 2020 – Dec 2022
Drag'n Survey
  • Fine-tuned BERT/CamemBERT for sentiment analysis across 16 languages, served via FastAPI (Flask → FastAPI migration, Dockerised)
  • Fine-tuned a handwritten OCR (Tesseract benchmarking, synthetic French dataset for easter2) and deployed an inference API
  • Experiment tracking with Neptune.ai; deployment on AWS Lambda + API Gateway
FastAPIBERT / NLPFine-tuningOCRAWS Lambda

Infrastructure & Dev Engineering

Jan – Jul 2022
Omnegy
  • Selenium scraping on AWS Lambda with GitLab CI/CD
  • Invoice parsing and ingestion pipeline into Salesforce CRM
AWS LambdaGitLab CI/CDWeb ScrapingSelenium

Interface Development

Oct – Dec 2021
BCA Expertise
  • Built a Streamlit web app for data scientists to analyse the output of an OCR pipeline
  • Refactored a monolithic codebase into gRPC microservices
StreamlitOCRgRPC

Data Processing Optimization

Jun – Sep 2021
Disney· Marne-la-Vallée
  • Migrated Excel-processing Python scripts from local to cloud to generate reports
  • Refactored, cleaned and deployed the scheduled (cron) jobs on AWS Batch
AWS BatchPythonRefactoring

Backend Automation Engineering

Nov 2020 – May 2021
Mobsuccess
  • Let non-technical users update e-commerce website data from Google Sheets via a FastAPI service
  • Added unit tests across several environments; Google Sheets API integration
FastAPIGoogle Sheets API

EEG Data Analysis

Feb – May 2020
Urgotech
  • Refactored NumPy code to optimise computation on EEG sensor data
  • Built a Streamlit app to visualise the data and debug the pipeline
StreamlitPythonNumPy

Classification Study

Sep 2019 – May 2020
WhiteLab Genomics
  • Feasibility study: unsupervised NLP to cluster scientific publications (scikit-learn, NLTK, spaCy)
  • Built a Scrapy pipeline to gather millions of scientific papers into DocumentDB (AWS)
NLPScrapyspaCyDocumentDBAWS

Teaching Experience

Reinforcement Learning courseDec 2022 – Mar 2023
AIX YNOV CAMPUS
Python & Machine Learning InstructorJan – Feb 2022
EFREI
NLP Introduction for PhD LinguistsJan 2022
Fénelon Sainte Marie
Introduction to PythonMay 2021
SIMPLON
Lead Formateur IANov 2018 – Apr 2020
SIMPLON

Education

Master of Neurosciences2013
UPMC · Paris
Bachelor of Physics & Chemistry2015
UPMC · Paris
Bachelor of Life Sciences2011
UPMC · Paris