Open to AI/ML, platform, backend, and product engineering conversations
Featured projects, current AI/product work, and backend systems.
Searchable publications, notes, and PDF previews inside the portfolio.
Persistent recruiter-focused AI assistant with source-linked answers.
Graduate training in Software Engineering & Intelligent Systems, hands-on experience at Radical AI, Hootsuite, and the University of Alberta, and a broader project archive spanning AI products, backend platforms, web systems, data workflows, and systems foundations.
The portfolio foregrounds real product AI, trustworthy ML evaluation, backend reliability, and delivery-quality software work rather than treating projects like a buried archive.
Built a user-facing AI assistant with GPT-4, structured prompting, personalization logic, feedback loops, and privacy-aware guardrails so career guidance felt fast, practical, and trustworthy.
Designed backend workflows for subscriptions and payouts with Stripe, Node.js, and Firebase, handling webhook events, retries, partial failures, state transitions, and idempotent processing.
Evaluated model brittleness under character-level and word-level adversarial perturbations, built reproducible evaluation scripts, and analyzed failure patterns across benchmark tasks.
Instead of forcing visitors to infer the story from scattered cards, the atlas makes the connections explicit: education, employers, flagship projects, research depth, and communication signal all reinforce each other.
The archive is broader, the current work is more visible, and the section is easier to read on large screens without the old overlap issues.
A user-facing AI assistant inside a web platform that delivered personalized, action-oriented career guidance with GPT-4, structured prompting, feedback loops, and privacy-aware guardrails.
Designed backend workflows for subscriptions and payouts in a SaaS context, integrating Stripe, webhook handling, retries, reconciliation logic, and idempotent processing.
Evaluated LLM robustness under character-level and word-level perturbations, automated perturbed datasets, and measured failure patterns across tasks such as AG News, SST-2, and MRPC.
A user-facing AI assistant inside a web platform that delivered personalized, action-oriented career guidance with GPT-4, structured prompting, feedback loops, and privacy-aware guardrails.
Designed backend workflows for subscriptions and payouts in a SaaS context, integrating Stripe, webhook handling, retries, reconciliation logic, and idempotent processing.
Evaluated LLM robustness under character-level and word-level perturbations, automated perturbed datasets, and measured failure patterns across tasks such as AG News, SST-2, and MRPC.
Created software and data workflows for biometric research, covering intake, cleaning, labeling, versioning, dashboards, and privacy-aware handling of sensitive data.
Built an ML classifier for phishing and malicious URL detection using feature engineering, preprocessing pipelines, and precision/recall/F1-driven evaluation.
Built a web and mobile coordination platform for humanitarian logistics with real-time tracking, Google Maps integration, authentication, and synchronized operational data.
Built a financial monitoring tool in Rust to ingest price data and compute technical and trend indicators efficiently over large datasets.
Engineered a QR-based system for registration, attendance tracking, real-time counts, exportable records, and organizer-friendly analytics workflows.
Implemented AVL Trees and Red-Black Trees in Rust with strong attention to correctness, invariants, safe memory usage, and thorough testing.
Designed and implemented an operating system simulator centered on process scheduling and memory-management concepts including Round Robin, FCFS, and virtual memory behavior.
This is designed like an interactive recruiter dossier: choose a role lens and the site updates the score, strongest evidence, interview probes, and next-step recommendation instantly.
Best when the role needs LLM product sense, prompt orchestration, evaluation instincts, and the ability to turn intelligent behavior into a usable software surface.
A concrete GPT-4 product that proves conversational UX, personalization, prompt design, and shipping judgment.
Robustness testing and validation thinking show that AI work here is not naive demo-building.
Node.js, APIs, and backend systems make the AI experience delivery-capable instead of model-only.
The personal layer now foregrounds the actual hiring signal: employers, graduate standing, conference participation, awards, and a broader trajectory across AI, backend, research, and teaching.
AI/ML and software engineer with graduate training in Software Engineering & Intelligent Systems, a completed MEng with a 4.0 GPA, and hands-on experience shipping intelligent product features, backend services, research tooling, and full-stack platforms across industry and academic environments.
Completed graduate studies with a 4.0 GPA while deepening ML systems, trustworthy AI, software verification, and intelligent systems foundations.
Shipped GPT-4 product capabilities, conversational flows, recommendation logic, logging, and quality loops for production-facing AI experiences.
Built research-grade tooling, reproducible data workflows, dashboards, documentation, and technical reporting for biometric research.
Worked across backend, billing systems, frontend components, and cloud functions with strong focus on reliability and maintainability.
Mentored students across software engineering, systems, databases, and programming while reinforcing technical communication and code-review discipline.
Provided long-term tutoring and mentorship in computing science, algorithms, systems, and software engineering fundamentals.
Designed and shipped GPT-4-powered product capabilities, built ReX end to end, added anti-fraud and validation layers, and improved latency, consistency, and usability of real-time AI interactions.
Built research tooling for biometric data collection, validation, dashboards, versioning, and reproducible analysis with privacy-aware handling of sensitive data.
Delivered production software across frontend, backend, and cloud functions with emphasis on Node.js services, Firebase, Stripe billing workflows, debugging, validation, and long-term maintainability.
Helped build, test, and deploy internal software across departments, connecting UI workflows, backend scripts, analytics, and operational reporting.
Recognized for academic excellence as part of the top 15% of eligible university students at the University of Alberta.
Affiliation with the Institute of Combinatorics and Its Applications, reflecting interest in combinatorics and discrete mathematics.
Affiliated with the International Communication Association through the Human-Machine Communication group focused on interaction with AI systems.
Selected participation and presentation history across research-oriented events in Canada, reinforcing research engagement beyond coursework.
Graduated from The British School of Kuwait with distinction, including A* A* A in Mathematics, Physics, and Further Mathematics plus 800/800 in SAT Math.
The site now surfaces the MEng 4.0, BSc 3.85, toolkit, coursework, and working style as first-class evidence rather than leaving them buried in the resume.
University of Alberta, completed in December 2025. Coursework included Software Construction and Verification, Robot Learning, Deep Learning in Computer Vision, Machine Learning System Engineering, and Blockchain Technologies.
University of Alberta, 2019-2024. Coursework across Artificial Intelligence, Software Engineering, Algorithms & Data Structures, Operating Systems, Networks, Databases, and Web Development.
Graduated with distinction, top 5% in the region, with A* A* A in Mathematics, Physics, and Further Mathematics plus 800/800 in SAT Mathematics.
Recognized for academic excellence and connected to combinatorics and human-machine communication communities through honours, affiliations, and research participation.
Ships intelligent product capabilities from prototype through production with a strong bias for reliability, clarity, and maintainability.
Moves comfortably across LLM orchestration, backend services, data workflows, evaluation, debugging, and user-facing product behavior.
Packages technical depth into interfaces, diagrams, and documentation that work for both technical and non-technical audiences.
Brings a mentoring and research-informed communication style shaped by teaching, tutoring, and university research tooling.
This experience is intentionally composed like a product: each layer answers a different question without losing visual restraint.
AI product work, backend reliability, ML evaluation, research tooling, and polished product delivery are all visible within the first few scrolls.
The portfolio surfaces the MEng 4.0, BSc 3.85, coursework, honours, and conference participation as real hiring evidence.
ReX, adversarial robustness, malicious URL detection, and research workflows create a stronger AI/ML story than a single flagship project would.
The expanded project archive now makes it easier to evaluate breadth without diluting the strongest current-work signals.
Navigation is simplified, spacing is calmer, cards wrap correctly, and the entire palette now leans into a sharper black-and-gold system.
The aesthetic is only believable if the underlying workflow is clear, rigorous, and built to survive production constraints.
Clarify who the system serves, what must remain reliable, and which interactions deserve safety, observability, and evaluation investment.
Design the service boundaries, data flow, prompt orchestration, validation layers, and interfaces that make the system understandable and durable.
Backend logic, AI behavior, UI polish, state handling, and documentation should land as one coherent product surface.
Logging, evaluation, testing, dashboards, and technical documentation extend the life of the system after launch.
The service layer is packaged for decision-makers: clear leverage, visible rigor, and high-conviction outcomes.
Design retrieval, memory, evaluation, and guardrail layers that turn prototypes into reliable AI products.
Translate ambiguous product requirements into clean interfaces, resilient services, and maintainable operating models.
Build polished interfaces that feel fast, clear, and trustworthy across desktop, mobile, and executive demos.
The terminal becomes a product cue: it suggests observability, confidence, and future extensibility while already remaining useful today.
Click nodes to inspect how the assistant, content layer, deployment strategy, and product surfaces fit together.
A practical AI product flow for delivering career guidance that feels fast, relevant, and trustworthy rather than generic.
Users interact through a guided conversational surface designed to produce actionable next steps rather than vague advice.
A backend flow focused on correctness, retries, idempotency, and the relationship between billing state and user entitlements.
Users interact with subscription and payout actions through a product experience that needs clear state and reliable feedback.
A reproducible research pipeline for sensitive biometric data that balances structure, privacy, and practical reporting.
Research requirements begin as study protocols that need to be translated into concrete software and data workflows.
These cards explain why the experience feels premium without inventing fake logos or made-up client quotes.
Projects move quickly, but the systems still stay legible. The experience is polished without hiding the engineering underneath.
Clear boundaries, strong narratives, and intentional diagrams make teams faster long after the first launch milestone.
Exploration is useful only when it becomes a decision. Experiments are packaged so founders and recruiters can actually evaluate them.
Posts deepen the story around latency, observability, engineering communication, and product-minded software delivery.
How to break response time into retrieval, reasoning, rendering, and follow-up loops without losing UX quality.
A field guide to evaluation sets, retrieval telemetry, and operator feedback layers for production retrieval systems.
Internal software shapes behavior. When the interface is calm and legible, systems become easier to trust.
The publications layer now carries more portfolio technical notes, demos, and PDF previews so the site feels closer to a real body of technical work.
A portfolio note on structured prompting, feedback loops, trust boundaries, and why user-facing AI assistants need product discipline as much as model capability.
A compact note connecting perturbation-based evaluation to real deployment questions around reliability, brittleness, and failure analysis.
A technical brief on subscription states, webhook handling, retries, entitlements, and why financial flows demand idempotent backend design.
The metrics and contribution terrain load from public GitHub data at runtime. If the API rate limit is hit, the interface falls back gracefully.
The contact layer now includes a real submission form, direct outreach paths, and a cleaner bridge from evaluation to conversation.
It is the quickest way to summarize fit, compare project areas, inspect research notes, and turn a general interest into a focused discussion.