Curriculum vitae
Gabin Maxime Nguegnang
AI Research Scientist & AI/ML Engineer
Munich, Germany · gmaxime@nguegnang-ai.com
Experience
AI Research Consultant: LLM reasoning and agentic AI evaluation
Outlier AI and Mercor · Remote, Germany
- Diagnose failure modes in multi-step reasoning of frontier LLMs on STEM problems, and write the corrections used as RLHF training data.
- Evaluate AI agent trajectories and function-calling precision on long-running tasks, separating agent errors from environment errors.
- Write the rubric criteria that grade the reward model, and lead peer review across a distributed team.
Doctoral research and teaching assistant
Ludwig Maximilian University of Munich · Munich, Germany
- Devised learning-rate conditions that guarantee convergence of stochastic gradient descent (SGD), the basis of reliable model training.
- Benchmarked 6 ensemble learners against 4 baselines in an explainable-AI anomaly-detection project on real diesel generator records, reaching F1 0.987 with LightGBM.
- Defended a disputed proof step with co-authors before the Springer Nature editorial board, through to acceptance.
- Taught postgraduate optimization, machine learning, deep learning and data science; co-supervised 1 master's student.
Doctoral research and teaching assistant
RWTH Aachen University · Aachen, Germany
- Derived learning-rate conditions for gradient descent on linear neural networks that do not decrease exponentially with network depth.
- Validated the theory and extended it to nonlinear architectures through large-scale PyTorch training on GPU clusters.
- Taught continuous optimization and mathematics of data science in a research group of 9 nationalities.
Machine learning intern
Group One Holding Company · Limbe, Cameroon
- Located sources of fuel loss across telecom base stations through exploratory data analysis, then trained a Random Forest (NSE 0.986) and shipped it with a monitoring dashboard, which helped account for 84,617 liters of fuel.
- Automated log ingestion (reporting time from days to seconds) and agreed with maintenance technicians which deviations the system should flag.
Education
PhD, Applied Mathematics: deep neural networks optimization
LMU Munich and RWTH Aachen University, Germany
Convergence analysis for training deep neural networks. Published in Springer Nature (2024) and SampTA 2023.
Co-operative MSc, Industrial Mathematics: machine learning
African Institute for Mathematical Sciences (AIMS), Limbe, Cameroon
Mastercard Foundation Scholars Program award, 2017.
MSc, Applied Mathematics: dynamical systems and modeling
University of Yaoundé I, Cameroon
Optimization and numerical analysis.
BSc Honours, Applied Mathematics
University of Douala, Cameroon
Publications
- Convergence of stochastic gradient descent for learning linear neural networks arXiv preprint, 2026
- Convergence of gradient descent for learning linear neural networks Springer Nature: Advances in Continuous and Discrete Models, 2024
- Analysis of stochastic gradient descent for learning linear neural networks SampTA 2023
- Anomaly detection in power generation plants using machine learning and neural networks Applied Artificial Intelligence, 2020
- Balancing Performance and Fairness in Explainable AI for Anomaly Detection in Distributed Power Plants Monitoring arXiv preprint, 2026
- Predicting fuel consumption in power generation plants using ML and neural networks ICECET 2021
Skills
Programming and data
Python · SQL · PyTorch · scikit-learn · pandas · FastAPI · Pydantic · PostgreSQL · DuckDB · pgvector · FAISS
LLMs, RAG, and AI agents
RAG (hybrid retrieval, reranking) · GraphRAG · AI agents · Multi-agent systems · LangGraph · MCP · Tool calling · Prompt engineering · Fine-tuning (LoRA, QLoRA) · DPO · RLHF · Hugging Face (Transformers, TRL, PEFT) · vLLM · LiteLLM
Knowledge graphs
Neo4j (Cypher) · RDF/OWL · SPARQL
Machine learning and statistics
Deep learning · NLP · Random Forest · Gradient boosting (LightGBM) · Anomaly detection · Explainable AI (SHAP) · Bootstrap confidence intervals · Cohen's κ · Optimization (OR-Tools, CVXPY)
Evaluation and MLOps
LLM evaluation (LLM-as-a-judge, ALCE) · Docker · Kubernetes · CI/CD (GitHub Actions) · Git · pytest · Azure (AKS) · AWS · Google Cloud · GPU clusters (SLURM)
Certifications
- Building GraphRAG Python MCP tools Neo4j GraphAcademy · Sep 2026
- Advanced agent coding Outlier AI · Jan 2026
- Model parallelism: building and deploying large neural networks NVIDIA · Oct 2025
- Generative AI with large language models DeepLearning.AI and AWS · Oct 2024
- Machine learning engineering for production (MLOps) DeepLearning.AI · 2024
Memberships and awards
- Munich Center for Machine Learning (MCML), member Jun 2024 – present
- Mastercard Foundation Scholars Program award Aug 2017
Languages
- French Mother tongue (C2)
- English Proficient (C1)
- German Professional working proficiency (B2+)