Data & Software Engineer
Data & Platform Engineer with nearly 5 years of production engineering experience, specializing in high-throughput distributed systems, real-time ETL pipelines, and cloud-native data platforms on AWS. Led end-to-end delivery of mission-critical platform migrations serving 20,000+ concurrent vehicles across 90,000 intersections, and built a data normalization platform producing ~2.36B Kafka messages/day powering all Analytics and Data Science workloads. Proven track record leading cross-functional teams, defining API contracts and data models, driving observability and CI/CD standards, and championing org-wide AI tooling adoption. Deep expertise in Kafka, Snowflake, Python, Kotlin, Java, AWS, and Kubernetes.
Here's a snapshot of the technologies I work with day-to-day.
Languages
Data & Streaming
Cloud & Infrastructure
Amazon Web Services
ECS Fargate · Lambda · DynamoDB
CDK · CloudFormation · Kinesis
Aurora / RDS · ECR · S3 · IAM · IoT
Kubernetes
Deployments · Services · HPA
Pods · ConfigMaps · Manifests
Docker · GKE · AWS EKS (familiar)
CI/CD & Observability
Databases & Additional Skills
Microservices & API Design
Nearly 5 years of production engineering at Global Traffic Technologies / Miovision, building systems that control traffic signals for emergency vehicles and transit buses across 90,000 intersections in North America.
Opticom Runtime Migration: Python / EC2 → Kotlin / Kubernetes
Skills: Kotlin, Kubernetes, AWS, Feature Flags, HPA
- Led zero-downtime re-architecture of the mission-critical EVP & TSP intersection-request system from GTT-owned AWS EC2 (Python) to Miovision AWS Kubernetes pods (Kotlin) — any failure is a live safety incident.
- Designed for 20,000+ concurrent emergency and transit vehicles, more than double the legacy system’s ceiling, using Kubernetes Deployments, Services, and Horizontal Pod Autoscaling.
- Implemented org-ID-based incremental customer cutover via the Feature Toggle Service, enabling individual organizations to be migrated independently during beta.
- Delivered ahead of schedule with zero production failures across the full cutover.
Opticom Analytics Platform — Data Normalization & Annotation Service
Skills: Python, Kafka, Snowflake, AWS ECS, Pydantic, Datadog, SQS
- De facto tech lead for a team of 3–4 engineers — broke down the epic, sequenced sprints in dependency order, pre-connected with SMEs across teams, and coordinated with PM to protect enterprise customer delivery timelines.
- Built a high-throughput annotation service processing 1,500 msg/sec with ~102.5M invocations/day, producing ~2.36B annotated Kafka messages/day — the single source of truth for all Opticom Analytics dashboards and Data Science workloads.
- Architected the real-time Kafka ETL backbone at 40M+ records/hour (~900M/day) with AVRO schema registry across distributed microservices.
- Streamed annotated output into Snowflake via Kafka connector, powering all downstream BI and Data Science consumption.
- Caught silent data loss during load testing via a Datadog SQS queue-depth dashboard — correctly diagnosed a resource-limit bottleneck instead of a bad date range, preventing false confidence in system capacity before production.
Emergency Vehicle Preemption (EVP)
Skills: Python, AWS Lambda, DynamoDB, Redis, IoTCore, SQS/SNS, CloudFormation
- Migrated four device families (210x, MP70, Cradlepoint, CEI) from Neptune to DynamoDB at 90,000-intersection scale, reducing operational costs and query complexity.
- Engineered a Redis cache with SQS/SNS-based invalidation in front of AWS Lambda device lookups, cutting P99 latency by 95% — from ~1,100ms to ~50ms — deployed across 90,000 live intersections.
- Wrote CloudFormation / Terraform IaC templates for all infrastructure, working alongside the deployment team to manage rollouts across environments.
Transit Signal Priority (TSP)
Skills: Python, AWS ECS Fargate, Lambda, CDK, Redis Pub/Sub, ECR, Pydantic
- Responsible for the full development and architecture of the TSP system — device support, data validation, real-time data handling, and cloud infrastructure via an event-driven architecture powered by Redis Pub/Sub.
- Implemented IoT device support for multiple hardware families, enabling seamless MQTT communication and data exchange for transit vehicles approaching intersections.
- Built Pydantic-based data validation to safeguard incoming IoT streams, plus a data aggregation service supporting playback and archival of live intersection requests — hosted on AWS ECS via CDK.
IoT Device Simulator
Skills: Python, Gradio, MQTT, AWS
- Identified that testing any Opticom feature required physical IoT hardware — expensive to ship, requiring vehicle installation and QA scheduling — and self-initiated a solution.
- Built a Python/Gradio simulator covering all device families (Sierra Wireless, Whelen lightbars, Cradlepoint, proprietary 2100 modems), each publishing to MQTT in their correct format (JSON or bytearray).
- Supports pre-recorded routes simulating realistic vehicle approaches to intersections for end-to-end EVP/TSP trigger validation from a laptop.
- Proposed, built, and merged into the main monorepo — adopted org-wide as the mandatory first sanity check for all new feature testing, with physical vehicle trips reserved for final QA only.
Centralized Feature Toggle Service
Skills: Python, AWS, API Design, Microservices
- Identified that each team managed feature flags independently via hardcoded values, env vars, or separate application stacks — creating redundant infrastructure and requiring redeploys for every new agency onboarding.
- Proposed and led development of a shared centralized Feature Toggle Service supporting both feature-level and org-level toggling granularity — adopted across all engineering teams.
- Eliminated redeploys for new customer agency rollouts, reducing onboarding to a pure configuration operation.
- Used as the org-ID cutover mechanism during the Opticom Kubernetes migration beta, enabling safe incremental customer migration.
API-based Video Sharing Platform on Google Cloud
Skills: Python, Flask, GKE, Cloud Functions, Cloud Storage, REST API, Kubernetes, Docker
- Architected and deployed RESTful APIs on Cloud Functions, integrating with Cloud MySQL and Google Cloud Storage.
- Developed a Python-Flask website hosted on Google Kubernetes Engine enabling video upload and viewing.
- Built container images, wrote YAML manifests, and deployed to GKE clusters, exposed to the public internet via a Kubernetes Service.
