Samuel Yeung

Senior software engineer with 10 years building production systems across the stack, currently a founding engineer at a venture-backed AI startup. Ship LLM and RAG infrastructure — retrieval pipelines, MCP tooling, citation-grounded agents — alongside the React platform and Kubernetes environment they run on. Comfortable owning a problem from stakeholder conversation through to cost-optimized production deployment.

Senior Software Engineer · Founding Engineer at Workorb AI (2023-01 – Present)

Founding engineer: shaped product direction directly with design partners and took the platform from zero to a venture-backed product generating six figures in revenue from early enterprise customers; now mentor junior engineers through design and code review.

• Architected the agentic orchestration system for an RFP-response product, backed by a RAG pipeline over OpenSearch k-NN embedding search — now the primary workflow for the majority of enterprise users and 80% of proposals produced.
• Designed a pluggable ingestion framework that delta-syncs terabytes of ACL-scoped content from Google Workspace, SharePoint (Microsoft Graph), and Slack into the RAG index, cutting new enterprise connectors from bespoke builds to half a day of work.
• Owned end-to-end delivery of a citation-grounded LLM chatbot backed by a custom MCP server exposing multiple retrieval tools (Java, Spring Boot, MySQL, GCP/Azure), saving users roughly 4 hours per week versus the previous search workflow.
• Prototyped a self-hosted alternative to a commercial LLM API for a latency-sensitive task: fine-tuned Qwen3-4B (4-bit QLoRA via Unsloth, single A100) on 2k human-labeled internal examples and served it under vLLM, reaching 70% answer accuracy against the incumbent model's 60% at materially lower cost per query.
• Scaffolded and maintain the React/TypeScript frontend platform; migrated Webpack to Vite and ESLint to Oxlint, cutting build time 70% and lint time 80%.
• Own the Kubernetes platform — right-sized workloads, automated dynamic resource allocation, and tuned autoscaling to cut compute spend 50%.
• Hardened the platform to SOC 2 compliance, remediating VAPT findings and closing security gaps that had been blocking enterprise deals.
• Cut query response times 90% by replacing ORM-generated queries with hand-tuned, index-aware SQL.

Senior Software Engineer at Cvent (2021-11 – 2023-01)

• Built serverless booking-reservation microservices (AWS CDK, GraphQL, Step Functions, SQS, DynamoDB, Lambda) on the critical path for millions of dollars in annual booking volume.
• Implemented room-availability search on Elasticsearch with Painless scripting, serving thousands of queries daily at roughly 200 ms latency.
• Volunteered to bring five unowned legacy services into compliance with a new dependency policy, replacing banned libraries and resolving breaking API changes across major-version upgrades — all shipped within a single sprint with no regressions.
• Cut CDK iteration time from 5+ minutes to under 1 by scripting stack delta updates in place of full synthesis, accelerating prototyping across the team.
• Drove service test coverage to 100% using JUnit, Testcontainers, and Mockito.

Senior Software Engineer at Aislelabs (2020-01 – 2021-10)

• Delivered a complete redesign of the frontend platform, devising an architecture that ran legacy and new UIs concurrently under constraints the team had considered unworkable; shipped a modernized interface and a 50% lift in user engagement.
• Rebuilt analytics on React over a legacy JSP stack and refactored the codebase behind it — tuned algorithms, introduced multithreading, and reworked database reads — cutting request times from minutes to seconds.
• Built a real-time marketing platform on an event-driven, queue-backed pipeline, sustaining peaks above 500 emails per second.
• Shipped a no-code visual workflow builder that let non-technical marketers design WiFi splash-page visitor journeys themselves, eliminating per-client custom development and anchoring a $100k+ contract with a major shopping-mall operator.
• Led a ground-up Shopify marketing integration, owning system design and directing task breakdown across a team of 3; adopted by dozens of enterprise customers.
• Extended Engage, an app-based indoor positioning system, with turn-by-turn navigation and a rebuilt localization layer that opened the product to non-English-speaking markets; integrated Stripe for card payments and recurring subscriptions.

Software Engineer at Aislelabs (2016-10 – 2019-12)

Software QA Automation Intern at IBM (2015 – 2015)

IT Analyst Intern at TD Bank (2014 – 2014)

Mobile Game Developer Intern at Clipwire Games (2014 – 2014)

Honours BSc in Computer Science – York University (2016-05)