Jason Chan

Aug 2026 to Dec 2026 5 months, full-time Boston, MA Kotlin, Android, Datadog

• Sole client engineer on an unreleased feature that gives players a passive, interactive reminder of DraftKings’ Android casino app between sessions; took it from proof of concept to an approved technical plan, settling its architecture and API contract with product and platform engineers; now building it in Kotlin behind a feature flag.
• Closed gaps in Datadog logging that were hiding failures in production, and fixed layout defects, across 3 DraftKings mobile apps (Casino, Sportsbook, Golden Nugget Online Casino) in 10 merged Kotlin fixes.

May 2026 to Jul 2026 3 months, full-time Foster City, CA

• Co-built a GenAI assistant that helped cut the rejection rate on Visa Treasury’s cross-border settlement onboarding form from 53% to under 15%; with one other intern, owned the chat widget, LLM orchestration, and validation, and delivered a working proof of concept in 8 weeks that later went to production.
• Designed the assistant deterministic-first, with regex and country-rule checks catching format and compliance errors before the LLM runs; the model sees only injected country data, with inputs allowlisted and internal IDs stripped.
• Built a rules validator for compliance attestations that resolves requirements per account tier and returns pass, fail, or human review, so ambiguous input goes to a person instead of a guessed verdict; verified with 127 unit tests.

Jan 2025 to Jun 2025 6 months, full-time Billerica, MA Java, Android, REST, JUnit, Jenkins

• Developed full-stack Android and Java features for FLIR’s uPoint Multi-Robot Control System, the controller operators use to run all 5 of FLIR’s ground robot lines in live field operations.
• Resolved 40+ prioritized engineering tickets in the 100K+ line codebase, including 8 shipped features, bug fixes, and system optimizations, on an Agile team with peer code review.
• Implemented REST endpoints for vehicle-to-operator state sync, and wrote JUnit suites, run in Jenkins CI, that raised test coverage on owned modules from 35% to 80%.

Sep 2025 Top 15 of 319 projects Python, TensorFlow, Flask, SocketIO

• Built a machine learning system that recognizes ASL gestures, serving classifications over Flask and SocketIO at sub-200ms inference latency; placed top 15 of 319 projects at HackMIT.
• Trained a CNN+LSTM on 90K+ frames across 250+ gestures to over 95% classification accuracy; added request batching and async processing to the server.

Apr 2024 Selected from 1,000+ applicants Built on-site at X HQ over a weekend Python, LLMs, X API, Chrome Extension

• Built a scam-post classifier for X that pulls a user’s timeline through the X API and flags likely scams with a few-shot LLM, surfacing candidates in a Chrome extension for the user to judge instead of hiding posts on their behalf.
• Wrote the Python pipeline that scored the classifier at 98% accuracy on 10K+ labeled examples; selected from 1,000+ applicants to build on-site at X HQ over a weekend.

Expected May 2027 Boston, MA GPA 3.8 / 4.0

Coursework

Algorithms & Data Structures, Object-Oriented Design, Operating Systems, Distributed Systems