Available for Projects

Jude
Mingay

Freelance AI/ML Consultant & Strategist

Beyond chatbots and LLM wrappers—I solve complex problems from scratch using deep expertise across the full ML spectrum. Computer vision, AutoML, time series, edge deployment, and custom architectures built on solid software engineering foundations.

3+ Years Experience

6 Languages Spoken

MSc AI Specialization

Start a Project View Experience

Based in

Freiburg, Germany

Currently

MSc @ Uni Freiburg

Focus

Computer Vision & DL

// What I Do

Expertise & Services

End-to-end machine learning solutions tailored to your business needs, from research to production deployment.

Computer Vision

Object detection, image segmentation, and custom vision models using state-of-the-art architectures like YOLO, RF-DETR, and transformers.

Object Detection Segmentation YOLO

AutoML & Optimization

Automated machine learning pipelines with hyperparameter optimization, neural architecture search (NAS), and meta-learning for rapid model development.

HPO NAS Meta-Learning

Edge & Inference

High-performance model deployment with TensorRT and Triton Inference Server. Optimized inference pipelines for production-grade throughput and latency.

TensorRT Triton ONNX

Industrial Protocols

Unique expertise in industrial automation protocols including IEC 61850, IEC 60870, and ICCP TASE.2 for energy and utility systems.

IEC 61850 SCADA Automation

Data Engineering

Custom annotation pipelines, synthetic data generation, and RAG systems with vector databases. End-to-end data infrastructure for ML.

Vector DBs Annotation Synthetic Data

Time Series & Forecasting

Predictive modeling for temporal data, anomaly detection, and demand forecasting. Deep learning and classical approaches for business intelligence.

Forecasting Anomaly Detection Temporal

// Featured Work

Projects

End-to-end ML solutions from concept to production deployment.

Client Project Apr 2025 — Ongoing

AIM

AI-Powered Football Coach

Architected and implemented the complete ML pipeline end-to-end—from building custom annotation tooling to production deployment. Extracts human-level coaching insights in real-time from just 2D video using multi-modal architectures.

RF-DETR RTMPose TensorRT GNN RAG Time Series

Visit Project

End-to-End

ML Pipeline

Internal Tool

AIM Annotation Suite

Multi-Modal Annotation

Full-stack annotation suite with SAM3 and RF-DETR auto-annotation, active learning, and on-demand GPU.

SAM3 RF-DETR Modal

Personal

FiscFox

German Tax Management

Privacy-first tax management with local AI, ML forecasting, and OCR receipts.

FastAPI TabPFN Ollama

Hackathon

StepMaker

CAD Work Instructions

Transforms CAD assemblies into multilingual work instructions with 3D viewer.

Next.js Three.js FastAPI

Hackathon

Greenvisor

Energy Efficiency AI

Cyclic RAG advisor for German home energy efficiency using knowledge graphs.

LangChain Neo4J FastAPI

Research

AutoML Text Classifier

Proxy Search Space HPO

Architecture-agnostic AutoML with custom proxy search spaces for optimizer selection. DEHB outperformed PriorBand via low-budget benchmarking.

DEHB NePS LSTM

Research

HPO4TabPFN

Scaling Down TabPFN

Discovered smaller, deeper models outperform larger configs. Muon optimizer dominated across all scales in TabArena benchmark.

TabPFN NePS Muon

Open Source LF Energy

FledgePower

Multi-Protocol Gateway

Developed gateway plugins for IEC 61850, TASE.2, and IEC 104 protocols. Enabled interoperability between power grid systems through protocol translation.

IEC 61850 TASE.2 IEC 104 C/C++

View Project

Protocols

// What They Say

Client Feedback

Building lasting relationships through quality delivery and clear communication.

Jude delivered an exceptional ML pipeline for AIM. His ability to architect complex systems—from custom annotation tooling to production deployment—while maintaining clear communication throughout was impressive. A true end-to-end engineer.

Big Picture Software Venture Studio • UK

Working with Jude on the AIM project has been a great collaboration. His deep technical expertise in computer vision and ML, combined with his practical engineering approach, helped us build a sophisticated AI coaching platform from the ground up.

Dristix Labs Development Partner • India

// Track Record

Professional Experience

Building ML systems that deliver real value

From research labs to production systems, I've worked across the full ML lifecycle. My unique background combines cutting-edge AI research with industrial software development, enabling me to bridge the gap between innovation and practical implementation.

Nov 2025 — Present

Senior Machine Learning Engineer (Freelance)

BIG PICTURE (SOFTWARE) LIMITED • Remote

Leading ML initiatives including object detection systems, transformer model deployment, and RAG implementation. Promoted from ML Engineer role after demonstrating technical leadership and delivery excellence.

Apr 2025 — Nov 2025

Machine Learning Engineer (Freelance)

BIG PICTURE (SOFTWARE) LIMITED • Remote

Developed custom computer vision models using YOLO and RF-DETR. Built end-to-end pipelines from data annotation to production deployment on GCP. Implemented TabPFN for tabular data predictions.

May 2024 — Mar 2025

Software Engineer

MZ Automation GmbH • Freiburg, Germany

Development and maintenance of libraries for IEC60870, IEC61850, and ICCP TASE.2 industrial communication protocols. Worked with C, C++, C#, Java in a test-driven development environment.

Jul 2023 — Apr 2024

Junior Software Engineer

Gridcore Labs Lda • Olhão, Portugal

Developed gateway frameworks for industrial communication protocols including IEC 61850, ICCP TASE.2, and IEC 60870. Foundation in industrial automation systems.

Mar 2023 — Jul 2023

Research Assistant

Universidade do Algarve • Faro, Portugal

Trained denoising and segmentation models for microscopy images of animal cells. First hands-on experience with deep learning and computer vision in a research setting.

// Technical Stack