Alejandro Rojas

5+ years writing and operating production software for a distributed data platform — a multi-cluster Trino / Starburst estate across AWS and on-prem, spanning source-level engineering, infrastructure automation, release engineering, and production reliability.

• ▸Engineer and operate a multi-cluster production Trino / Starburst estate across AWS and on-prem, providing federated data access across a heterogeneous data ecosystem.
• ▸Java, Python, SQL, and shell development across the Trino codebase — proxy / integration components, query-performance engineering, and code-level query / platform behavior troubleshooting.
• ▸Cut mean time to recovery from roughly 3 hours to roughly 1 hour through production troubleshooting, fleet-scale observability, incident response, and runbook / automation work that removes repeat failure classes.
• ▸Ship weekly zero-downtime deployments across the production Trino / Starburst estate using GitHub Actions CI/CD, with pre-release query replay and validation to catch regressions before rollout.
• ▸Implemented Snowflake on-behalf-of authentication through Trino so users reach data with their own identities, with Privacera-based data-layer access control integrated with OAuth / identity.
• ▸Supported a heterogeneous data ecosystem — hands-on Databricks / Unity Catalog integration and troubleshooting (connectivity, auth / access, query compatibility), plus Snowflake, Spark, Tableau, Python clients / connectors, SSIS, and Iceberg across authentication, connectivity, query behavior, and platform compatibility.
• ▸Kept AWS and on-prem infrastructure as declarative, reviewable Terraform + Ansible; GitHub Actions CI/CD; Kubernetes / Helm for platform workloads; and S3-related migration work.
• ▸Contributed to a professional agentic security / compliance platform: repository / context ingestion, multi-step agent workflows, structured findings, deterministic guardrail / policy evaluation, risk / impact analysis, and remediation guidance.

Technical project · Personal AI infrastructure

Designed and operate heterogeneous local LLM inference infrastructure across multiple NVIDIA GPU systems, serving LLMs behind OpenAI-compatible APIs with SGLang, vLLM, and llama.cpp. Hands-on work in quantization, long-context / KV-cache tuning, speculative decoding, and containerized serving.

Drexel University: Master of Science in Software Engineering

• ▸Graduate Certificate in Software Architecture
• ▸Graduate Certificate in Artificial Intelligence & Machine Learning
• ▸Graduate Certificate in Software Security