At AMD, I worked on out-of-band management and infrastructure around the development of the MI300X, MI400X GPU accelerator series. I utilized Terraform and Ansible to provision and manage 50+ Linux VMs, ensuring proper disk mounts, partitions, and user management. I also resolved a Docker container synchronization issue due to a race condition between the loading of kernel modules and starting of systemd services.

During my time at SAP, I worked on Just Ask and SAP Analytics Cloud focus on NLQ (Natural Language Querying). I delivered multiple features & user stories end-to-end. Notable projects included the optimal granularity algorithm and the unbooked cascading effect. The primarily challenge was making sure that the overhead was minimal (< 100ms) leveraging selector caching mechanisms and the implementation was robust enough for over 34K+ customers. I also worked on some separate topics such as microservice resilience (circuit breaker, time/rate limiters) to reduce RTT and CI/CD pipelines with Jenkins for deployment / testing.

During 2024W1 and 2024W2, I mentored over 150 students collectively through personal office hours, guiding them software development practices and object-oriented design. I also contributed to question development through Docker and Github Codespaces, besides running labs and assisting with lectures. Answered over 200+ questions on the discussion board.

During my first SWE internship, I worked on ETL pipelines, processing Kafka-streamed data from multiple systems on a CRON-based schedule. I also made sure that preprocessing was efficient and strongly advocated for eventual consistency between the database layer and the execution tracker, adhering to the BASE principles, given the differing upsert mechanisms (SQL vs GraphQL). I also worked on the Elasticsearch implementation to index and support querying across 3000+ SKUs.

Wrote market making and order book algorithms for a trading simulator platform, supporting Sauder's trading competition. Leveraged Websockets to broadcast prices in real-time to multiple clients, and Redis ZSET for efficient leaderboard implementation.

Managed the Serverless backend for UBC's Largest Technology Club. Developed endpoints to interface with DynamoDB, and leveraged GSI (Global Secondary Index) for efficient querying patterns across 20+ events and 300+ members. Utilized Cloudformation for resource management across multiple development / production environments.

Mentored 5 product managers and 20 product developers to develop projects for UBC's Google Developer Student Club.

Spearheaded the resume review platform for GDSC's annual flagship event, leveraging AWS S3 and RDS to store and manage resumes and reviews for 100+ attendees and 10+ partners, also leveraged the localStorage API to avoid redundant submissions.

Developed an experiment framework for UBC VCL, utilized Express, jsPsych, jQuery, and d3.js to create graph visualizer app based on random researchers' input conditions, supporting over 15+ graphs and 35+ features.

Architected ennovateubc.org, a platform for an entrepreneurial competition, utilizing Next JS, MongoDB for rapid prototyping. I also implemented pagination and debouncing for the team page to improve user experience for teachers across 100+ local schools, and prevent server overload.

Developed bbrwanda.org, a platform for a 20+ year old non-profit organization designed to foster Rwanda's sustainable development initiatives through practical learning and action.

Developed ubctc.ca, a platform for UBC's Tennis Circle, a student-run organization that provides opportunities for students to learn and improve their tennis skills.

Achieved the Dean's Scholar, Faculty of Science International Scholarship, and Charles and Jane Banks Award.

Mentored middle-high school students programming with Python and Java, helping them build foundational coding skills and problem-solving abilities.