Mohsen Hadian

Data Scientist with 8+ years building end-to-end ML/AI systems across healthcare, engineering, and education. Recent focus on LLM-powered retrieval-augmented generation (RAG) systems and multi-agent frameworks for clinical/policy decision support, computer vision and medical image analysis (classification/segmentation), and integration of large multi-site clinical and EHR datasets on cloud-scale infrastructure (AWS, Azure, GCP/BigQuery). Strong collaborator with clinicians and hospital teams; translates research into reproducible, containerized, production-ready code.

Data Scientist (Remote) - Toronto Western Hospital - Toronto, ON

(2023-01)

• LLM & RAG systems: designed and deployed PolicyPal AI, a retrieval-augmented generation assistant for policy Q&A, and CARE-Concussion, a multi-agent LLM framework supporting concussion-care decision workflows — evaluating retrieval quality with RAGAS and iterating on prompt/agent design based on failure-mode analysis.
• Clinical NLP: built a clinical Named Entity Recognition pipeline (seqeval-evaluated) to extract structured entities from unstructured patient PDFs, turning previously unusable free-text records into structured fields for downstream analytics.
• Computer vision / medical imaging: led ML/AI analyses across the ONDRI multi-site neurodegenerative disease cohort, integrating MRI-derived imaging features (cortical thickness, volume, free water) with clinical and symptom data using CNN- and U-Net-based feature extraction, feeding into concussion and neurodegeneration risk models.
• Big data / dataset integration: built semantic column-mapping pipelines using sentence-transformer embeddings to harmonize longitudinal, multi-site clinical datasets with inconsistent schemas at scale on GCP/BigQuery — used across the WSIB/WCB return-to-work concussion risk stratification work and the ICES concussion sex-differences population study.
• Rigor & delivery: applied site-aware cross-validation, calibration, and subgroup error analysis; produced clinician-facing dashboards and documented model cards/runbooks for reproducibility and hand-off.
• Governance: established data-handling standards (de-identification, access controls) aligned with Canadian privacy expectations.

Sessional Faculty — Machine Learning & Business Analytics - University Canada West - Vancouver, BC

(2022-01 - 2025-01)

• Teach ML/analytics (Python/Excel/Tableau); mentor student projects end-to-end (EDA → modeling → reporting).
• Emphasize reproducible code, validation, and communication of model behavior to non-technical stakeholders.

ML Consultant — Reinforcement Learning for Cancer Therapy - Tarbiat Modares University - Tehran, Iran (Remote)

(2022-01 - 2023-01)

• Designed RL/optimization pipelines to explore dosing policies coordinating chemo- and immunotherapies; ran simulation studies to compare policies.
• Simulation results showed the combined chemotherapy + anti-angiogenic policy reduced tumor cell counts by ~65% versus either treatment alone, with the RL-tuned MPC reaching up to 90% optimality in constrained scenarios.

Data Scientist (Part-time) - ImiRoad - Victoria, BC

(2021-01 - 2023-01)

• Built data processing/feature pipelines and recommendation components to power product personalization (profile-driven logic and decision trees).

Data Scientist (Part-time) - Teachingate - Saskatoon, SK

(2020-01 - 2023-01)

• Developed NLP-driven models for automated feedback (English writing support); delivered dashboards and reports for stakeholders.

Machine Learning Engineer - University of Saskatchewan - Saskatoon, SK

(2019-01 - 2023-01)

• Computer vision: designed a robust predictive controller and obstacle-detection algorithm for vision-guided robotics, and led perception/control ML deployed in MATLAB, ROS, Simulink, and Python.
• Built end-to-end experimentation pipelines (data capture → training/evaluation → deployment) and performed ablation/tuning studies.
• Improved mobile robot response time by 250% using RNN-based control models.

AI Engineer - Petroleum University of Technology - Ahvaz, Iran

(2015-01 - 2019-01)

• Applied heuristic optimization and ML (e.g., SVM, clustering) to industrial and physical systems; collaborated with domain experts on model validation.

Data Scientist - Khuzestan Cochlear Implant Center - Ahvaz, Iran

(2016-01 - 2018-01)

• Analyzed clinical outcomes, designed studies, and contributed to quantitative/qualitative mixed-methods research.

Data Scientist - Jondishapur University of Medical Sciences - Ahvaz, Iran

(2014-01 - 2016-01)

• Big data / predictive modeling: built deep neural network models on large public-health datasets to forecast mortality and comorbidity risks (HIV, HCV, addiction).
• Designed intervention simulations to support policy decisions in public health.

University Lecturer & Lab Instructor - Petroleum University of Technology - Ahvaz, Iran

(2015-01 - 2019-01)

• Taught programming and engineering labs; supervised student projects in data analysis, control, and optimization.

PhD - Biomedical Engineering - University of Saskatchewan (2019-01 - 2023-01)