Software engineer focused on DevOps, platform engineering, infrastructure, and distributed systems.
My work spans production SRE at FindingFive, Terraform-based provisioning, Docker environments, and versioned CI/CD pipelines. I also build distributed applications and investigate request routing for AI inference.
At Stony Brook, I’m studying Computer Architecture, Analysis of Algorithms, and Database Systems—connecting instruction execution, algorithmic complexity, and storage design to the performance of the systems I build.
A multi-team academic platform needed consistent infrastructure and dependable releases. Manual environment setup made development and deployment slow and difficult to reproduce.
My contribution
Automated provisioning with Terraform and Docker, standardized Windows and macOS developer setup, and built versioned release workflows with validation and traceability.
The result
Reduced deployment from approximately one day to under ten minutes. Supported nine reproducible releases across three development teams, with improved environment consistency and rollback readiness.
02 / InfrastructureIn production
mydininghall.com
A campus dining idea that became a service for real students. Real-time menus, background workers, and an observable cloud platform.
2 campusesOswego + Vanderbilt
CAMPUS → CLOUD
m↗mydininghallGood food. Less guesswork.
Today’s menuOswego ↗
◒Fresh, in one placeMenus · nutrition · dining hours✓
APIRedisWorkers
DjangoReact / Next.jsPostgreSQLRedisCeleryAWS
Behind the build
The challenge
University dining information was difficult to navigate. The platform brings menus, nutrition information, and dining hall hours together in a responsive experience.
My contribution
Worked across application infrastructure, cloud deployment, databases, caching, and asynchronous processing. Celery and Redis automate menu collection with fallback handling; Prometheus, Grafana, Loki, and Umami provide operational visibility.
The result
Expanded from a SUNY Oswego capstone into a multi-tenant service for Oswego and Vanderbilt. The portfolio records 1.2k monthly visitors and a $20k investment, with a white-label architecture for adding universities.
Can the signals we collect make the systems we serve faster? Exploring tail latency and request routing across AI inference backends.
3 strategiesone instrumented testbed
SIGNALS → DECISIONS
requestrouter
vLLM 01vLLM 02vLLM 03
latency · load · utilization ↩
GovLLMPrometheusGrafanaAWS EKSTerraform
Behind the build
The question
Does observability-aware routing reduce tail latency in AI inference, and which signals matter most?
The approach
Built a testbed with multiple inference backends to compare round-robin, least-connections, and observability-aware routing under load. Metrics connect routing decisions with latency, throughput, and resource utilization.
The context
Spring 2026 independent study advised by Professor Douglas Lea, connecting infrastructure, router development, and load experiments with analysis of model-serving performance.
Making a behavioral research platform more observable, easier to deploy, and faster to recover.
Contributions
Replaced nightly pull-based deployments with push-based Bitbucket Pipelines and rollback support.
Introduced New Relic APM, distributed tracing, and synthetic monitoring; identified approximately two seconds of latency for Korean users accessing US servers.
Reduced incidents by 70% through proactive monitoring and improved MTTR by 30% with automated runbooks and systemd scripts.
Established versioned Python packaging and webhook-based Slack deployment notifications. The platform serves 1K+ monthly active users.
Building repeatable delivery workflows and helping development teams ship with confidence.
Contributions
IBM capstone project (January–May 2026): developed versioned GitHub Actions and Terraform deployment workflows, container builds, and pull-request quality gates.
CSC480 build coordinator: implemented four multi-stage pipelines for PR validation, Docker builds, GHCR publishing, and versioned deployments.
Coordinated QA, usability, and stakeholders as the CSC480 project's sole DevOps engineer, supporting a 20-person agile team.
GitHub ActionsTerraformAWSDockerGHCR
Open sourceAug 2025 — Present
Open Source Contributor
Falcon · Dokploy · GitLab / CodePath AI301
Improving the tools other engineers rely on, one reviewed contribution at a time.
Contributions
Falcon: fixed suffixed-responder routing bugs and streamlined cross-platform Python wheel CI using cibuildwheel.
Dokploy: contributed profile image uploads, API/CLI metadata customization, and an SSH IP-address parsing fix.
Contributed to public GitLab issues through CodePath's AI Corps program.
PythonCI/CDCode review
Community leadershipAug 2024 — May 2025
Technical Lead
Google Developer Groups · SUNY Oswego
Helping a student developer community learn through collaboration and practical engineering.
Contributions
Coordinated technical activities and collaborative initiatives, supported students learning development tools, and encouraged peer learning and teamwork.
MentorshipTechnical leadership
Teaching & mentorshipTA: Jan 2023 — Dec 2024
Teaching Assistant & Academic Tutor
SUNY Oswego
Making difficult ideas approachable through patient, practical technical communication.
Contributions
Supported systems programming and data structures courses through office hours and grading.
Tutored Calculus I and Data Structures and Algorithms, accumulating approximately 200 hours of tutoring experience.
Data structuresSystems programmingCalculus
InternshipFeb 2022 — Jul 2022
Machine Learning Software Engineer Intern
Juncture
Building an NLP-powered resume parsing MVP for candidate and job matching.
Contributions
Developed and deployed a spaCy-based parsing pipeline for local talent acquisition clients; reported 86% parsing accuracy.
Built Selenium collection infrastructure for 10,000+ LinkedIn profiles to support skill analysis and training datasets.
PythonspaCySeleniumNLP
Engineering is a people practice, too. Campus Technology Services help desk work and Resident Assistant responsibilities taught me to troubleshoot carefully, communicate clearly, and take ownership.
03
Research
Research & systems thinking.
AI INFERENCE · INDEPENDENT STUDYSpring 2026
Can better visibility make inference faster?
Exploring whether observability-aware request routing can reduce tail latency across AI inference backends—and which signals make the difference.
Outside the application layer—and outside work—I’m usually learning how something works. Sometimes that’s a Linux VM in my homelab. Sometimes it’s a camera, a new city, or a problem a student is working through.
01PhotographyPeople, places & architecture02Homelab & hardwareLearning by running real systems03Teaching & communityApproximately 200 hours of tutoring
Get in touch
I’m interested in SRE, platform, infrastructure, backend, and AI systems opportunities, as well as research collaborations in distributed systems and infrastructure.