Julian
Quezada

Software Engineer

New York, NYOpen to new roles

I build infrastructure and tools that scale. Previously at LinkedIn (Spark Infrastructure) and KBRA (Data Engineering). I like hard problems and clean systems.

View Work ↓GitHub →

Impact

By the numbers.

Spark Infrastructure · LinkedIn

~$1M saved
per year

Built a compute optimization pipeline that analyzed Spark DAGs, detected redundant cache opportunities, and auto-notified repo owners — estimated savings validated with manager and skip-level.

Performance · Spark History Server

10 min
→ instant

Implemented caching for the Spark History Server. 1,000+ engineers in the data org went from waiting to instant job debugging.

Data Migration · KBRA

39 tables
Zero data loss

Migrated ~20GB from legacy SQL Server to AWS S3 and Snowflake. Redesigned ingestion from FTP to S3. Validated with row counts and deduplication.

Daily Impact · Cyclops

6 analysts
every day

Built Cyclops — an internal Python Dash app that replaced a manual bond securities review process. CUSIP validation, mismatch detection, and automated email reports in one tool.

Experience

Where I've built things.

LinkedIn
Junior Software Engineer
Spark Infrastructure
Feb 2022 – May 2024

  • Built a Grafana dashboard to monitor Spark deployments across all pods and environments — used weekly by on-call engineers to instantly identify failing jobs.

    60–100 jobs tracked per podweekly on-call usage

  • Implemented caching for the Spark History Server, cutting job UI load time from ~10 minutes to near-instant across the data org.

    ~10 min → near-instant1,000+ engineers benefiting

  • Built a compute optimization pipeline that analyzed Spark DAGs to detect redundant computations, mapped them to source repos, and auto-notified owners.

    ~$1M/year in compute savings
PythonJavaScalaApache SparkHDFSGrafanaSQL
KBRA
Data Engineer
Cyclops — 2-person team
May 2024 – Mar 2026

  • Built Cyclops — an internal Python Dash app used daily by the content management team to surface bond securities mismatches and automate CUSIP validation.

    6 daily active usersreplaced manual review process

  • Migrated 39 tables (~20GB) from legacy SQL Server to AWS S3 and Snowflake with zero data loss. Redesigned ingestion pipeline from FTP to S3.

    39 tables~20GB migratedzero data loss

  • Built a CI/CD validation tool using Playwright to automate post-deployment browser testing — catching version mismatches without manual verification.

  • Served as internal AI advocate. Evaluated Codex, Claude Code, and GitHub Copilot for company adoption and presented recommendations to leadership.

PythonSQLPython DashPlaywrightSnowflakeAWS S3GitLab CI/CD

Projects

Things I've built.

📊

AI Cost Analytics

Cost Estimator

OpenRouter cost analytics dashboard. Tracks spend, token usage, and model costs across providers with charts and breakdowns.

↗ Live
🎵

Oct 2020 – Jan 2021

Harmony Social

Music-focused social platform with Twilio and Spotify APIs. React SPA served by Express with a Node.js background worker.

🍓

Feb – Mar 2021

Strawberry Sound

Shazam-like song identification tool. Recognized 2,000+ songs in two weeks using Web APIs and AuD Music Recognition.

🏆

2022 – 2023

Hackathon Wins

1st at Citi Bank (Google Cloud Vision ML). 2nd at Valtech × Create Labs (GPT-3 resume summarizer). Two wins, two different stacks.

Skills

Tech Stack.

Languages

PythonJavaScalaSQLTypeScriptJavaScriptCC++Rust

Data & Storage

Apache SparkHDFSSnowflakeAWS S3MongoDBPostgreSQLMySQLTrino

Cloud & Infra

AWS EKSAWS Secrets ManagerGoogle CloudKubernetesDockerGitLab CI/CD

Frameworks

ReactNext.jsNode.jsExpressFlaskReact Native

Tools

GrafanaPython DashPlaywrightAzkabanArtifactory

Specialties

Spark DAG OptimizationData MigrationObservabilityCI/CD Pipelines

Contact

Let's work
together.

Emailjulianq.js@gmail.com→LinkedInlinkedin.com/in/haxagonus→GitHubgithub.com/HaxagonusD→