Adrian H.
Senior Machine Learning Engineer / Data Scientist
Experience
Oct 2023 - Present
2 years 3 monthsSenior Machine Learning Engineer / Data Scientist
Artefact
- Designed an evaluation workflow for language model responses with automated scoring logic and structured human review steps, giving more stable assessments across several internal tasks.
- Added retrieval-based prototypes using Python, PyTorch and vector indexing to support context-aware reasoning over internal documents; included checks for reproducibility and fixed-seed variants.
- Set up repeatable benchmark datasets with documented scoring criteria so different prompt revisions and fine-tuned versions could be compared in a consistent way.
- Analyzed failure cases in generated outputs through exploratory data investigation, which shaped later dataset selections for fine-tuning efforts.
- Created small end-to-end analytical tasks used to test model reasoning chains, covering data preparation, statistical summaries and simple predictive components so research teams could check model weaknesses in realistic workflows.
- Worked closely with product and engineering groups to turn experiment findings into staged rollout plans.
- Reduced evaluation iteration time noticeably by refactoring some of the Python utilities and adding lightweight tooling for traceability.
Mar 2020 - Sep 2023
3 years 7 monthsSenior AI/ML Data Scientist
Miquido
- Designed A/B testing plans and statistical validation steps to measure business impact of new features that relied on predictive models.
- Prototyped NLP, recommendation and structured-data models in PyTorch and Hugging Face; compared several model families against baseline approaches using deterministic workflows.
- Created and maintained feature engineering pipelines in Python and SQL to support recurring dataset updates and reproducible experiments.
- Ran exploratory studies to diagnose changes in model outputs and identify sampling shifts or data quality issues.
- Standardized experiment tracking using MLflow, making comparisons between model versions easier for both analysts and engineers.
- Worked with engineering teams to refine post-deployment monitoring signals and added checks for data integrity that were missing before.
- Added a set of analytical problem scenarios used internally to evaluate reasoning in model candidates; these covered data cleaning, feature creation, basic forecasting and interpretation steps.
- Mentored less experienced teammates on experiment framing, documentation practices and communicating results clearly.
Nov 2017 - Feb 2020
2 years 4 monthsMachine Learning Engineer / Applied Research Engineer
PELTARION
- Compared multiple modeling strategies using scikit-learn, TensorFlow, and PyTorch to understand trade-offs in predictive stability and data requirements.
- Built dataset preparation pipelines and feature transformations in Python and SQL to enable controlled experimentation.
- Used Jupyter-based analyses to present findings to product and engineering partners, helping narrow research scope to high-value modeling paths.
- Added internal scripts for logging experiment runs and visualizing performance trends, making research iteration more consistent.
Oct 2016 - Nov 2017
1 year 2 monthsData Scientist
Seldon
- Performed exploratory analysis and hypothesis testing to surface behavioral patterns and support planning discussions.
- Created predictive scoring models in Python and validated performance using retrospective evaluation and controlled experiments when feasible.
Sep 2015 - Oct 2016
1 year 2 monthsSoftware Engineer
Neoteric
- Wrote Python utilities for data extraction and transformation used by analytics and reporting groups.
- Coordinated with analysts to ensure consistent data formatting and interpretation across pipeline steps.
Summary
Senior Machine Learning Engineer with over 10 years working across data science, research prototyping and model evaluation for practical product use. Experienced in statistical modeling, experimentation, and designing evaluation frameworks that support clear decisions. Comfortable working in exploratory research as well as setting up experiments for real-world rollouts. My work over the years has focused on careful reasoning, structured analysis, and incremental improvements grounded in evidence. Recently, my focus has included language model evaluation, retrieval workflows, and problem design involving multi-step analytical reasoning.
Languages
English
AdvancedEducation
Sep 2020 - Aug 2022
University of Essex
Master of Science in Computer Science · Computer Science · Colchester, United Kingdom
Sep 2011 - May 2015
Hanoi University of Science and Technology
Bachelor of Science in Computer Science · Computer Science · Hanoi, Viet Nam
Need a freelancer? Find your match in seconds.
Try FRATCH GPT More actions
Similar Freelancers
Discover other experts with similar qualifications and experience
- English
- Deutsch
2025 © FRATCH.IO GmbH. All rights reserved.