Evidence before adjectives
A result should name its dataset, protocol, metric, and limitation. If the artifact is missing, the claim is not promoted.
About / Working approach
Mathematical care, practical engineering.
I work at the boundary between modelling and systems: understanding what a model can support, then building the data, evaluation, and software path that makes that evidence useful.
01Current focus
Applied ML with research depth
Mohd Zamin QuadriAI/ML EngineerMunich, Germany
My recent work centres on uncertainty-aware graph neural network surrogates, calibration, conformal prediction, and selective review. Alongside that research, I build reference ML pipelines and grounded AI service prototypes to exercise the operational side of model development.
My Master's thesis, Uncertainty Quantification for Machine Learning Models in Transportation Policy Analysis, was submitted at Technical University of Munich. The available record states: Master's thesis submitted. No later education status is published.
I am based in Munich, Germany. Open to full-time Machine Learning and Applied AI roles, including roles across Machine Learning Engineering, Applied AI, reliable ML, scientific computing, GNNs, MLOps, and data/AI engineering.
02Disciplines
The kinds of work, not the order of them
Grouped by discipline rather than run as a timeline, because with five roles the order is the only thing a sequence adds and it is the thing a reader over-reads. These approved titles establish the public record without inferring dates, private client details, or unsupported impact figures.
Applied AI systems
Retrieval, verification and the engineering around a model rather than the model alone: storage that suits the question, and a service whose output can be checked against its own evidence.
AI Engineer (Working Student)
BP-IT Consulting & Solutions GmbH / Munich
Built verification workflows for a multilingual legal knowledge platform using relational, vector, and graph storage.
Numerical methods and scientific visualization
The discipline underneath the rest of this portfolio. Numerical behaviour, and making a computed result legible to the person who has to judge it.
Student Research Assistant, Numerical Methods and Scientific Visualization
Technical University of Munich
Student Research Assistant / Programming and Visualization
Technical University of Munich
Machine learning research
Estimating a quantity that is not directly measurable, and being careful about what the estimate can support. The battery work is the earliest instance of the question the thesis later formalised.
Summer Research Intern, Machine Learning for Li-ion Battery State Estimation
IISER Bhopal
Data and workflow engineering
Moving data between systems that were not designed to talk to each other, which is where most of the practical difficulty in an ML pipeline actually lives.
Intern, Programming of Workflows and Linking of Databases
AUDI AG
03Education and academic roles
Mathematics carried into computation
The approved public record includes TUM student research assistant titles in programming, visualization, and numerical methods. No duties beyond those titles are inferred.
Technical University of Munich
M.Sc. program: Mathematics in Science and Engineering
Master's thesis submitted
Aligarh Muslim University
B.Sc. (Hons.) Mathematics
04Principles
How I approach technical work
Reliability is a system property
Model quality includes calibration, failure modes, data contracts, monitoring boundaries, and the human decision around a prediction.
Small, inspectable interfaces
I prefer explicit pipeline stages, typed service contracts, deterministic fixtures, and tests that make assumptions visible.
Scope is part of the result
Research, coursework, prototypes, synthetic demonstrations, and deployed systems answer different questions and should be labelled accordingly.
05Capabilities with proof
What I can contribute
Machine Learning & Research
Graph models, uncertainty quantification, calibration, experiments, and evidence-led evaluation
Generative AI & Retrieval
Grounded retrieval, vector search, citation-aware responses, and local model integration
Backend & Data Systems
Typed APIs, data validation, reusable transformations, service boundaries, and traceable artifacts
MLOps & Engineering
Experiment tracking, promotion gates, model registries, containers, CI, and testable workflows
Data Science & Analytics
Scientific computing, sensitivity analysis, time series, diagnostics, and reproducible reporting
Next
The clearest version of this is the work itself.
Each case study states the problem, the evidence, and the limitation. The repository index shows the same standard applied to smaller projects.