Backend engineer on Snap Map's team (400M+ MAU). I own the Go/Python ETL and geospatial systems that ingest, reconcile, and rank data from 10+ third-party providers (Google, Yext, TripAdvisor, Ticketmaster, SeatGeek).

What I've shipped:

• Architected the Places ETL in Go - the single source of truth for every place search, profile, and story served to 400M+ MAU.

• Recovered 4.1M corrupted place records (822K names, 1.1M categories, 3.9M geometries) after a provider-priority misconfiguration wiped curated attributes. Used BigQuery forensics + authored a reusable bulk-update script, now the team's standard playbook.

• Grew daily tracked POIs 80K → 200K+ (2.5x) and lifted total events on Snap Map 35% by building an end-to-end POI ETL in Go pulling real-time data from Ticketmaster, SeatGeek, and Gametime with geospatial matching + priority scoring.

• Shipped Smart POI Name Cleaning by integrating a 4B-parameter LLM (Qwen3-4B) into the POI pipeline via Snap's AGI Inference Service - auto-cleaning ~500 cluttered event titles per ingestion cycle.

• Drove a 29% lift in place-tagged Spotlight Snaps by designing the backend for "Spotlight Posting Hints" in the Send-To-Spotlight cell.

• Increased Ticketmaster POIs 3.5x and SeatGeek POIs 2x via concordance-based place matching - onboarded SeatGeek (~470K venues) and added Ticketmaster venue-ID lookups, hitting 89% POI-to-Place match.

• Owned pipeline observability: Grafana dashboards + alerting; authored the POI Improvement Roadmap (15+ proposals, 2 shipped as P0 same quarter); completed the HLS → STMS migration.

Graduate TA and Grader for CS-GY 9223: Cloud Computing and Big Data Systems - a masters-level course covering distributed systems, Hadoop, Spark, and cloud platforms.

• Supported 100+ graduate students through office hours, code review, and grading - achieving 95% student retention through engaging instruction and personalized feedback on distributed systems and big-data assignments.

Cloud Engineering team at a Fortune 500 financial services firm. I worked on reusable infrastructure tooling and security automation used by platform teams across the org.

• Cut deployment time 10% and improved security posture 15% by building reusable Terraform modules and enforcing security policies with CloudFormation Guard and Cloud Custodian across the AWS estate.

• Reduced manual incident-response work 30% by building a Flask-based containerized service that auto-creates ServiceNow tickets for critical process failures - now embedded in the team's on-call workflow.

• Reduced cloud infrastructure costs 5% via automated exception handling and resource optimization across CI/CD pipelines.