Kevin Dong

This is basically just a copy of my resume (link here) and my LinkedIn. Not much to see here otherwise.

Reach out [email protected] if you want to get in touch. My GitHub's @kevindong.

Experience

Still on Metrics Index!

Software Engineer on the Metrics Index team.

Worked on real-time metrics data ingestion and query service used on all recent metrics data (Mork), including virtually all alerting/monitoring evaluations. Ingests trillions of datapoints a day and responds to queries for said datapoints at high scale and availability. Service is primarily built in Rust.
Cross-functionally collaborated with teams across Metrics, Query, and Alerting orgs to design, implement, and deliver improved alerting/monitoring reliability to end-customers during internal incidents
Proposed and implemented CPU throttler that effectively eliminated an entire class of previously frequent (1-2 times per week) incidents and source of on call burden
Designed and achieved metrics-wide adoption of a shared, programmatic understanding of query prioritization. Previously, each team would locally decide priorities, which frequently clashed between services and lead to excess query failures.

Developer on the Latin American Feeds & PFCOs team. Team is responsible for connecting to Latin American stock exchanges and certain non-stock-exchange-firms, parsing data, and sending structured data to downstream Bloomberg teams. Work is primarily in C++ and Python.
Modernized (i.e. rewrite or incrementally improve) legacy market data parsers to increase reliability and meet internal quality/behavioral standards

Built a distributed, fault-tolerant, and horizontally scalable data deduplication Java microservice capable of processing >25,000 events/second in collaboration with a data engineering team; service leveraged Kafka transactions and partitions to shard incoming events into the microservice hosted on Kubernetes
Reduced processing times by 90% for certain critical operations in Praetor (an internal Go gRPC microservice for A/B testing and feature flagging) while improving reliability and safety; evangelized Praetor to other teams and orgs within the company
Built a blackbox monitor in Go to measure the reliability and responsiveness of Praetor and to output metrics to Prometheus/Grafana with new alerting functionality
Held primary responsibility for monitoring and responding to issues arising in production for the company's main traffic-serving system (along with all systems/services owned by my team) during on-call shifts; also provided support for other engineers

iOS developer in consumer banking business
Prototyped an iOS app in Swift that interfaces with internal APIs/microservices
Sourced and utilized an open source library for interacting with a Go server via WebSockets

Developer on the Chicago Police Department team using C# with ASP.NET MVC and Oracle SQL
Rewrote multiple data warehouses search applications (originally in Visual Basic) using ASP.NET MVC
Analyzed, architected, and rewrote web application (originally in Perl) for police lineups. Police lineups are a series of mugshots shown to a victim/witness of a crime wherein the victim/witness will point out a suspect. This process, if conducted fairly, is admissible evidence in court. Work includes database layer, web UI, everything in between, and a robust audit log system.

Roughly listed in order from best to good (1 = best; 8 = good).