Looks like you have JavaScript disabled. For the full Twine experience, you will need to re-enable it.

Machine Learning and Computer Vision Engineer with 5+ years of experience building Python/C++ systems for real-time video analytics, biometric recognition, and biomedical signal processing. Strong end-to-end capability across data collection/annotation, preprocessing, model development, evaluation, latency optimization, and production integration. Experienced collaborating with engineering and public-sector stakeholders, designing modular and reproducible pipelines others can run and extend. Proven ability to maintain accuracy under real-world conditions such as occlusion and illumination changes, with practical MLOps skills including Docker, FastAPI, and CI/CD.…Machine Learning and Computer Vision Engineer with 5+ years of experience building Python/C++ systems for real-time video analytics, biometric recognition, and biomedical signal processing. Strong end-to-end capability across data collection/annotation, preprocessing, model development, evaluation, latency optimization, and production integration. Experienced collaborating with engineering and public-sector stakeholders, designing modular and reproducible pipelines others can run and extend. Proven ability to maintain accuracy under real-world conditions such as occlusion and illumination changes, with practical MLOps skills including Docker, FastAPI, and CI/CD.

Milad Parvan

Programmer, Data Scientist, AI Engineer, +2





Machine Learning and Computer Vision Engineer with 5+ years of experience building Python/C++ systems for real-time video analytics, biometric recognition, and biomedical signal processing. Strong end-to-end capability across data collection/annotation, preprocessing, model development, evaluation, latency optimization, and production integration. Experienced collaborating with engineering and public-sector stakeholders, designing modular and reproducible pipelines others can run and extend. Proven ability to maintain accuracy under real-world conditions such as occlusion and illumination changes, with practical MLOps skills including Docker, FastAPI, and CI/CD.…Machine Learning and Computer Vision Engineer with 5+ years of experience building Python/C++ systems for real-time video analytics, biometric recognition, and biomedical signal processing. Strong end-to-end capability across data collection/annotation, preprocessing, model development, evaluation, latency optimization, and production integration. Experienced collaborating with engineering and public-sector stakeholders, designing modular and reproducible pipelines others can run and extend. Proven ability to maintain accuracy under real-world conditions such as occlusion and illumination changes, with practical MLOps skills including Docker, FastAPI, and CI/CD.

Available to hire

Machine Learning and Computer Vision Engineer with 5+ years of experience building Python/C++ systems for real-time video analytics, biometric recognition, and biomedical signal processing. Strong end-to-end capability across data collection/annotation, preprocessing, model development, evaluation, latency optimization, and production integration.

Experienced collaborating with engineering and public-sector stakeholders, designing modular and reproducible pipelines others can run and extend. Proven ability to maintain accuracy under real-world conditions such as occlusion and illumination changes, with practical MLOps skills including Docker, FastAPI, and CI/CD.

Skills

Experience Level

Expert

Expert

Expert

Expert

Expert

Expert

Expert

Expert

Expert

Expert

Intermediate

Intermediate

Beginner

Beginner

Language

English

Fluent

Turkish

Fluent

Persian

Fluent

Dutch

Beginner

Work Experience

Computer Vision Research Collaborator at Vrije Universiteit Brussel (VUB)

February 1, 2026 - Present

Developing and evaluating a computer-vision pipeline for eye-region detection, pupil localization and size estimation, temporal tracking, and gaze estimation from RGB video. Collecting and organizing video data to develop and evaluate pedestrian attention-analysis methods, testing reliability across distance/illumination/head motion/image quality, and documenting conditions where pupil measurement degrades. Preparing technical documentation covering data collection, software implementation, evaluation methodology, and experimental findings for a visual-attention analysis application.

Independent Research (ECG biometrics bench) at Self-directed

February 1, 2025 - February 1, 2026

Completed and released ECG-biometrics-bench as a reproducible single-author benchmarking study for ECG biometrics. Published a preprint on arXiv; under review at Computers in Biology and Medicine.

Independent Researcher at Self-directed

February 1, 2025 - February 1, 2026

Completed and released ECG-biometrics-bench as a reproducible benchmarking study for ECG biometrics. Authored a preprint on arXiv (under review at Computers in Biology and Medicine) focused on configurable benchmarking procedures and reproducible evaluation for ECG-biometric methods.

Machine Learning Research Engineer at University of Milan

March 1, 2024 - February 1, 2025

Engineered a modular Python benchmarking framework integrating 30+ iris-recognition algorithms behind reusable interfaces to improve reproducibility and reduce effort for evaluating new methods. Developed and maintained ECG-biometrics-bench with configurable preprocessing, feature extraction, training, and evaluation stages. Improved ECG recognition performance (~20%) by modeling longitudinal user-aging effects and adding metric-driven evaluation workflows. Designed repeatable experiment workflows and documentation to support model comparison across datasets and configurations.

AI/ML Engineer (Consultant) at Road Traffic Injury Research Center

January 1, 2020 - January 1, 2023

Developed a real-time computer-vision perception pipeline for vehicle detection, tracking, speed estimation, and inter-vehicle distance measurement from roadside camera streams. Designed multi-stage video-processing workflows including frame preprocessing, localization, temporal tracking, geometric estimation, validation, and results generation. Curated and annotated ~100 hours of video; implemented reproducible training/validation and failure-case analysis pipelines. Improved robustness under rain, shadows, illumination changes, and partial occlusion, achieving ~6% improvement versus GPS-derived ground truth. Optimized for low-latency execution and operational reliability and presented findings to government stakeholders.

Research Assistant (Part-time) at University of Tabriz

February 1, 2019 - February 1, 2024

Developed a reusable EEG/BCI experimentation framework for signal preprocessing, feature extraction, model training, and comparative evaluation of ML, DL, and reinforcement-learning methods. Refactored data-processing and experimental workflows, reducing execution time by ~40% and enabling near-real-time analysis. Designed an anomaly-detection solution with monitoring and alerting that contributed to ~70% reduction in faulty production output. Produced documentation and a laboratory handbook and supported junior researchers in implementing/evaluating ML pipelines.