Teaching & mentoring
Teaching & mentoring
I teach quantitative methods by connecting assumptions, derivation, implementation, diagnostics, and interpretation. My formal teaching spans undergraduate and graduate mathematics, while my current mentoring emphasizes simulation design, identifiability, reproducible computation, and careful interpretation of model-based evidence.
Teaching philosophy
I ask students to derive a method, implement it carefully, test its behavior, and explain what the resulting evidence does and does not support.
Teaching experience
Courses taught, assisted, or supported.
My formal teaching spans undergraduate and graduate mathematics, including algebra, calculus, linear algebra, differential equations, and numerical analysis.
West Virginia University
- Fall 2021
Algebra with Applications
Instructor
- Spring 2021
Multivariable Calculus
Teaching Assistant
- Spring 2020
Applied Linear Algebra
Teaching Assistant
- Spring 2019
Numerical Analysis
Teaching Assistant
- Fall 2018
Advanced Linear Algebra
Teaching Assistant
- Fall 2018
Engineers’ Calculus
Recitation Leader
Shiraz University of Technology
- Spring 2009
Advanced Engineering Mathematics
Instructor
- Fall 2008
Differential Equations
Teaching Assistant
Mentoring
Mentoring through research questions, code review, and interpretation.
Research mentoring
As a Trainee Mentor and Hackathon Co-Organizer for the MISM Center of Excellence Summer Trainee Program at Duke in 2026, I advised trainees on simulation design, parameter identifiability, and code review.
Additional mentoring and academic service
I have also served as an External Graduate Student Panelist for an NSF REU, an Invited Graduate Judge at WVU research symposia, a mentor to three M.Sc. students, and the founder and president of the WVU SIAM Student Chapter.
Future curricular development
Graduate courses prepared for future development.
These are proposed courses, not courses I have already taught. Each course connects statistical or computational methods with specific health or biomedical applications.
- Statistical Learning for Health DataProposed · Introductory graduate · advanced undergraduate
- Causal Inference for Observational Health RecordsProposed · Master's
- Uncertainty-Aware Clinical ML for Perioperative & Critical-Care DataProposed · Doctoral · flagship
- Bayesian Methods for Biomedical NetworksProposed · Doctoral seminar
- Optimal Transport for Biomedical Data AlignmentProposed · Doctoral seminar
Proposed courses · syllabus details
Expandable syllabus outlines
Each proposed outline specifies the course purpose, intended audience, prerequisites, learning objectives, module sequence, assignments, grading, final project, datasets, tools, readings, and place within the broader teaching program.
01 Introductory graduate · advanced undergraduate Statistical Learning for Health Data Penalized regression, tree ensembles, conditional-independence testing, and survival analysis, with each method paired with a public clinical dataset and a reproducible analysis.
- Catalog description
- A hands-on introduction to statistical learning for students entering data science from outside mathematics, taught through health data. Each method is approached through derivation, working code, and a complete analysis of a public clinical dataset.
- Why it matters
- Health data science requires practitioners who can fit a model and defend the choices that produced it. From the first week, calibration, subgroup performance, and label noise are treated as parts of the modeling problem rather than as afterthoughts.
- Target audience
- First-year graduate students and advanced undergraduates in data science, biomedical informatics, statistics, or clinical fields entering quantitative work.
- Prerequisites
- One course each in probability/statistics and programming (Python); linear algebra recommended.
- Learning objectives
- Fit and evaluate supervised models on clinical data and reason about bias–variance and regularization.
- Test conditional independence and perform principled feature selection.
- Analyze censored outcomes with survival models.
- Produce a reproducible analysis another student can rerun.
- Module schedule
- Reproducible workflows: notebooks, environments, data manifests.
- Regression & loss functions.
- Penalized regression: ridge, lasso, elastic net.
- Classification, logistic regression & calibration.
- Resampling, cross-validation & leakage.
- Tree ensembles: random forests, gradient boosting.
- Conditional-independence testing & feature selection.
- Survival analysis: Kaplan–Meier, Cox models.
- Evaluation for health: AUC pitfalls, subgroup metrics.
- Missing data & interpretability (permutation importance, PDP).
- Assignments
- Weekly labs, each on a public clinical dataset; a mid-term "why is this AUC misleading?" critique.
- Grading
- 45% weekly labs · 15% critique · 30% final project · 10% participation.
- Final project
- Choose a public clinical dataset, define a prediction task, fit at least two model families, evaluate with calibration and subgroup analysis, and submit a reproducible notebook with a short written report.
- Datasets
- MIMIC-IV derived tables, a public sepsis cohort, UCI clinical datasets, SEER (for survival).
- Tools
- Python, scikit-learn, lifelines, pandas, Jupyter, Git.
- Readings
- James et al., ISLR; Hastie, Tibshirani & Friedman, ESL (selected); selected clinical-ML method papers.
- Program fit
- This course provides the statistical and computational foundation for the causal-inference and uncertainty-aware courses that follow.
02 Master's Causal Inference for Observational Health Records Targeted maximum likelihood, instrumental variables, and sensitivity analysis taught alongside the practical constraints of EHR data, including missingness, irregular sampling, label leakage, and governance.
- Catalog description
- Methods for estimating causal effects from observational health records, taught alongside the data-generating processes that can invalidate naive estimates.
- Why it matters
- Clinical decisions often require causal questions, including for which patients does this intervention help? EHR data, however, are confounded, irregular, and incomplete. The course pairs principled estimation with explicit sensitivity analysis.
- Target audience
- Master's and doctoral students in data science, biostatistics, and biomedical informatics; clinical fellows with quantitative training.
- Prerequisites
- Regression and probability; a prior ML or statistics course; Python.
- Learning objectives
- Formalize causal questions with potential outcomes and DAGs.
- Estimate effects with IPW, doubly-robust methods, TMLE, and instrumental variables.
- Conduct sensitivity analysis for unmeasured confounding.
- Diagnose EHR-specific biases and interpret treatment-effect heterogeneity.
- Module schedule
- Causal questions & potential outcomes.
- DAGs & identification.
- Confounding, adjustment & propensity scores.
- IPW & doubly-robust estimation.
- Targeted maximum likelihood (TMLE).
- Instrumental variables.
- Time-varying treatments & g-methods.
- Heterogeneous treatment effects.
- EHR realities: missingness, irregular sampling, label leakage.
- Sensitivity analysis (E-values); governance, fairness & transportability.
- Assignments
- Problem sets on identification and estimation; an EHR "bias audit" of a published observational claim.
- Grading
- 40% problem sets · 20% bias audit · 30% final project · 10% participation.
- Final project
- Pose a causal question on a public EHR cohort, justify identification, estimate with a doubly-robust method, run a sensitivity analysis, and report with clinical interpretation.
- Datasets
- MIMIC-IV, eICU, public claims subsets.
- Tools
- Python, DoWhy, EconML, scikit-learn; R optional.
- Readings
- Hernán & Robins, Causal Inference: What If; Pearl, Causality (selected); Peters, Janzing & Schölkopf (selected); TMLE method papers.
- Program fit
- The methodological core of the causal-phenotyping and treatment-effect direction in my teaching program.
03 Doctoral · flagship Uncertainty-Aware Clinical ML for Perioperative & Critical-Care Data Three units on probabilistic calibration, distribution shift, and subgroup auditing across the data sources that shape clinical AI: EHR, laboratory measurements, medications, high-frequency physiological waveforms, imaging-derived variables, notes, and biomolecular profiles.
- Catalog description
- A doctoral course on building and auditing clinical machine-learning systems that report uncertainty explicitly, organized around calibration, distribution shift, and subgroup performance.
- Why it matters
- Models that perform well on average may still fail during induction, decompensation, weaning, or recovery. Evaluation must therefore examine uncertainty and failure patterns, not only average accuracy.
- Target audience
- Doctoral students in data science, statistics, biomedical informatics, and computer science; postdocs; clinical fellows in research electives.
- Prerequisites
- A machine-learning course; probability/statistics; Python. Causal inference helpful.
- Learning objectives
- Calibrate and evaluate probabilistic predictions with proper scoring rules.
- Apply conformal prediction and quantify uncertainty.
- Detect and quantify distribution shift across time and sites.
- Audit subgroup performance, fairness, and instance-level harm.
- Combine supervised prediction, phenotyping, representation learning, and causal inference on multimodal clinical data.
- Module schedule
- Unit I: Calibration.
- Probabilistic prediction & proper scoring.
- Calibration methods & reliability diagrams.
- Conformal prediction.
- Uncertainty quantification.
- Covariate & label shift.
- Shift detection & monitoring.
- Domain adaptation & optimal transport.
- Temporal & multi-site shift.
- Subgroup metrics & fairness.
- Instance-level error & harm scoring.
- Governance & reproducibility.
- Assignments
- Labs on calibration, conformal sets, shift detection, and subgroup audits; a paper-replication assignment (reproduce one figure from a clinical AI paper).
- Grading
- 40% labs · 30% final project · 20% paper replication & critique · 10% participation.
- Final project
- Define a cohort (justify inclusion/exclusion), construct features, train and evaluate a model, and write a report combining quantitative results with clinical interpretation. Submit reproducible code and a data manifest another student can rerun end-to-end on a SLURM cluster or containerized workstation.
- Datasets
- MIMIC-IV and MIMIC-CXR (PhysioNet-credentialed), eICU; physiologic-waveform subsets.
- Tools
- Python, PyTorch, scikit-learn, MAPIE (conformal), Docker, SLURM/HPC.
- Readings
- Guo et al. (calibration); Angelopoulos & Bates (conformal prediction); clinical distribution-shift literature; subgroup/fairness auditing papers.
- Program fit
- This course integrates the principal themes of my teaching program and aligns directly with my MIT reliability work and Duke postdoctoral research.
04 Doctoral seminar Bayesian Methods for Biomedical Networks Ensemble structure learning, MCMC over graph spaces, and bootstrap assessment of stability for gene-regulatory and immunological networks, with edge-level uncertainty reported explicitly.
- Catalog description
- A research seminar on Bayesian network inference for biomedical systems, drawing on ensemble structure learning and stability analysis.
- Why it matters
- Biomedical network analysis requires reproducible structure estimates with explicit uncertainty, rather than a single graph produced by one algorithm and one data sample.
- Target audience
- Doctoral students in data science, statistics, and computational biology; methodologically oriented postdocs.
- Prerequisites
- Probability and a graphical-models or ML course; Python; some Bayesian statistics.
- Learning objectives
- Represent and score Bayesian networks.
- Run constraint- and score-based structure learning.
- Explore graph posteriors with MCMC.
- Assess edge stability via bootstrapping and report calibrated uncertainty.
- Module schedule
- Graphical models & d-separation.
- Conditional-independence testing with multiple-hypothesis correction.
- Score-based search.
- Constraint-based algorithms.
- Ensemble / hybrid structure learning.
- MCMC over graph spaces.
- Bootstrap stability & edge uncertainty.
- Deconvolution & network-immunology case study.
- Evaluation & benchmarks.
- Project presentations.
- Assignments
- Implement and apply a structure-learning pipeline; lead a seminar critique of a published network-inference paper.
- Grading
- 35% implementation · 25% critiques & seminar leadership · 30% project · 10% participation.
- Final project
- Infer a network from a real biomedical dataset with calibrated edge stability, and report the structure together with its uncertainty.
- Datasets
- Public gene-expression (TCGA subsets); immunology / cytometry references.
- Tools
- Python, bnlearn, PyMC, causalnex, TETRAD, NetworkX.
- Readings
- Koller & Friedman (selected); Latifizadeh et al., BaMANI (arXiv, 2025); Klinke et al., Nature Communications (2022).
- Program fit
- The seminar develops the causal-structure methodology used in BaMANI and supports the broader causal-phenotyping direction.
05 Doctoral seminar Optimal Transport for Biomedical Data Alignment Entropic and unbalanced optimal transport, with attention to relaxed-marginal formulations for cross-species and cross-site alignment of biomedical measurements.
- Catalog description
- A seminar on optimal transport for aligning biomedical distributions across instruments, sites, and species, emphasizing unbalanced and relaxed-marginal methods.
- Why it matters
- Measurements and models can shift across sites, instruments, and species. Distribution alignment should therefore report both transferred structure and unmatched mass.
- Target audience
- Doctoral students in data science, statistics, computational biology, and computer science.
- Prerequisites
- Linear algebra, probability, and optimization basics; Python.
- Learning objectives
- Formulate optimal transport (Monge and Kantorovich).
- Compute entropic OT with the Sinkhorn algorithm.
- Apply unbalanced and relaxed-marginal OT.
- Use Wasserstein distances in machine-learning pipelines for alignment.
- Module schedule
- OT foundations.
- Kantorovich duality.
- Entropic regularization & Sinkhorn.
- Wasserstein distances & barycenters.
- Unbalanced OT.
- Relaxed-marginal formulations & unmatched mass.
- OT for domain adaptation.
- Gaussian-mixture embeddings & graph clustering.
- Cross-species / cross-site case study.
- Project presentations.
- Assignments
- Implement Sinkhorn; apply unbalanced OT to an alignment task; lead a critique of an OT-in-biology paper.
- Grading
- 35% implementation · 25% critiques & seminar leadership · 30% project · 10% participation.
- Final project
- Align two real biomedical datasets with (unbalanced) optimal transport, reporting transfer performance and unmatched mass.
- Datasets
- Flow-cytometry references; multi-site expression subsets.
- Tools
- Python, POT (Python Optimal Transport), scikit-learn, scanpy, Leiden/igraph.
- Readings
- Peyré & Cuturi, Computational Optimal Transport; Cuturi (Sinkhorn); Chizat et al. (unbalanced OT); Latifizadeh et al., OT-RMC (in preparation).
- Program fit
- The seminar develops the optimal-transport methods used in my cross-dataset harmonization work and supports multi-site phenotyping applications.