Experience

Research appointments

Trustworthy AI Research Lab, Toronto Metropolitan University · Toronto, Canada · June 2026–present
Supervisor: Prof. Reza Samavi

Develop certified-training objectives using neural-network bound propagation and selective SOCP-derived regularization.
Study LP, SOCP, and SDP relaxations and evaluate clean, adversarial, and certified performance in reproducible experiments.
Collaborate on certified robustness and score-based quickest change detection with researchers at Princeton University and Augusta University.

Trustworthy AI Research Lab, Toronto Metropolitan University · Toronto, Canada · September 2021–May 2026
Supervisor: Prof. Reza Samavi

Developed neural-network robustness certificates using convex relaxations and cascading verification.
Studied selective lifting and sparse constraints to improve certificate quality while managing computation.
Produced research software, benchmark evaluations, manuscripts, and presentations; assisted with student mentorship.

Analog Integrated Circuit Design Laboratory, University of Tehran · Tehran, Iran · 2017–2020
Supervisor: Prof. Shahin Jafarabadi Ashtiani

Modeled, optimized, and simulated inductive wireless power and data transfer systems.
Contributed to a deep-brain-stimulator project through design and evaluation of amplifiers, coils, and rectifiers.

jamais51 Technologies Inc. and Toronto Metropolitan University · Toronto, Canada · June 2026–present

Formalize threat models, integrity conditions, and evidence requirements for an independent settlement verifier.
Develop adversarial tests for structured record corruption and characterize observable failure modes and information limits.

ParagraphAI · Toronto, Canada · June–September 2026

Led product and data work, coordinating developers and domain experts around measurable requirements and validated releases.
Designed controlled experiments and analyzed engagement, funnels, campaigns, and revenue to estimate effects and guide product decisions.
Configured AI-agent workflows for analytics, content generation, and business automation; refined prompts and used structured evaluation to assess accuracy, consistency, and task performance.
On SpeechLP, a project within ParagraphAI, collaborated with speech-language pathologists and developers on product planning, content verification, and experiments for speech-therapy games.

Research affiliation: Vector Faculty Affiliate Researcher. Research on trustworthy and robust machine learning.

See Research for the methods and projects and CV for a consolidated summary.