SYS://KERRYGHAN.RELOT v2.6.1
UPLINK UTC · 02:14:08
Initializing portfolio

DATAtoDEPLOY

Ingest
Train
Ship
Monitor
TRANSFER · NODE 01 000%
// dataset ready · entering portfolio
SYS://KERRYGHAN.RELOT v2.6.1
Open to work · France UTC · 02:14:08
NODE // 01 SIG // LOCKED LAT // 48.85N
Open to work · France

I build the machinery that turns data into decisions.

Data Analysis · AI Engineering · MLOps

AI engineer working at the seam between research and production — designing data pipelines, training models, and shipping the infrastructure that keeps them reliable.

02 // PROFILE

About

Operator at the seam of research and production — building the pipelines and monitoring that let models stay honest after launch.

// PORTRAIT.SIGRES 1024×768 · 60FPS

AI engineer and MSc Artificial Intelligence graduate, working at the intersection of data, modeling, and infrastructure — the work that turns research into something a team can actually rely on.

My focus spans the full lifecycle: analyzing the data, training and evaluating models with Python, PyTorch and Scikit‑learn, then wrapping them in pipelines and monitoring so they hold up in production.

I care about systems that stay honest after launch — reproducible training, observable serving, and metrics people trust enough to make decisions on.

// Formation

  • MSc Artificial Intelligence
  • BSc Computer Science
  • Certifications in Machine Learning & Cloud Computing
03 // STACK

Skills

The toolkit I reach for across the model lifecycle — from first analysis to production monitoring.

01

Machine Learning

PythonPyTorchTensorFlowScikit-learnNLPDeep Learning
02

Data Analysis

PandasNumPyMatplotlibSeabornJupyterSQL
03

MLOps & Deployment

DockerKubernetesCI/CDMLflowAirflow
04

Data Engineering

SparkKafkaETLData PipelinesPostgreSQL
05

Cloud

AWSGCPLinuxTerraform
06

Programming

PythonGitFastAPIBashReact
04 // BUILDS

Projects

A few systems I've designed, shipped, and kept running in production.

Realtime Churn Forecasting

Gradient-boosted ensemble serving 40M predictions a day at sub-80ms latency, with automated drift detection.

XGBoostFeastKafka

RAG Copilot for Analysts

Retrieval system over 2M internal documents, gated by a custom faithfulness evaluation harness on every release.

LangGraphpgvectorFastAPI

Feature Platform Migration

Moved 300+ offline features into a versioned online store, driving training–serving skew to near zero.

dbtAirflowRedis
05 // UPLINK

Contact

Have a project in mind, or a problem you're trying to make tractable? I'd be glad to hear about it.

Let's stay in touch

Reach out and I'll get back to you within a couple of working days. Short messages welcome — we can take it from there.

Phone

+33 1 23 45 67 89

Location

Paris, France

// Availability

Currently open to freelance projects and full-time roles. Comfortable working remotely or on-site across the EU.

Send me a message

Fill out the form below and I'll reply as soon as I can.