I’m Gokalp, a platform and applied AI engineer in San Diego. I work with teams to understand their systems, simplify everyday work, and put AI to practical use.
San Diego, California17+ years across infrastructure, platforms, and applied AI
Things I’ve built.
Projects from my platform work and my own exploration of AI tools for engineers.
Atlas
Exploring system architecture.
I built Atlas to help engineers find their way through a large system: what connects, what depends on what, and who owns it.
Sample architecture
The thinking behind Atlas
Why I built it.I built Atlas to help engineers explore dependencies, trace paths, and understand ownership. Its AI ingestion workflow uses Cortex code intelligence. An internal rollout at Arcules is being prepared to address missing architecture documentation and diagrams.
How I keep views grounded.AI proposes what to show, and a validator checks the proposal against the architecture catalog. Connections come from the system’s architecture model. Each view also identifies the entities and relationships it has summarized or left out.
Cortex
Code context for AI agents.
I built Cortex to give AI agents the code, dependencies, and past decisions they need to work in a repository.
Find the right code
Trace dependencies
Assess change impact
Carry decisions forward
How I built Cortex
Cortex combines syntax-aware parsing across 25 languages, hybrid semantic and keyword retrieval, and symbol-level dependency analysis. It serves targeted context through MCP and indexes locally.
Incremental refresh and explicit freshness and confidence signals help distinguish stronger evidence from inferred relationships. Persistent decisions and checkpoints support work across sessions.
Newman Operator
Testing in the team’s workflow.
I worked with developers to bring repeatable API and browser tests into their existing development and CI workflows.
Define the test
→
Trigger the run
→
See the result
One operator, two kinds of tests
I built a Kubernetes operator supporting Newman API tests and Playwright browser tests through separate custom resource types, with trigger-based execution, CI integration, and notifications.
I partnered with developers on authentication containers so test execution fits the applications and environments the teams actually use.
Planwright
Keeping track of agent work.
I built Planwright to follow what an agent has done, why it made a decision, and whether the work has actually been verified.
Sous project · requirements and progress
Sous project · task details and verification evidence
A personal project, built through use.Planwright grew out of my work on Cortex. I added versioned plans, acceptance criteria, and verification evidence so I could review an agent’s work and trace its decisions.
Completion and verification, tracked separately.During Atlas development, Planwright exposed tests that passed without the required implementation and components that were never connected to the application. Those failures shaped stronger checks.
Where I’ve worked.
My experience spans customer consulting, infrastructure, and leading a platform team. Here are a few results and the roles behind them.
¼ the cost.
I built an observability platform for 200+ microservices at Arcules, guided by developer feedback and usage analysis.
5× faster.
My Python migration tooling at Intuit moved 20 million artifacts and 100 TB of images, outperforming the native tools.
Zero downtime.
At Lytx, I migrated mission-critical IoT messaging to AWS Managed Kafka without downtime.
9 engineers.
Distributed platform team I lead at Arcules, with office hours and feedback loops that shape the roadmap.
2022 — Present
Arcules
Site Reliability Engineering Lead
I work with development teams to turn operational friction into platform capabilities. Developer surveys and usage analysis shaped an observability platform on VictoriaMetrics that now serves more than 200 microservices at a quarter of the previous cost. The same conversations led to Newman Operator for API and browser testing, a Prometheus library teams adopt with a few lines of code, and an ArgoCD plugin that lets them ship dashboards and alerts alongside their applications.
More recently I connected Git and Jira to AI-assisted release notes, GitLab releases, and Slack. I lead nine distributed engineers and keep developer office hours so the platform follows what teams actually need.
2020 — 2022
Lytx
Senior Systems Engineer
At Lytx, I worked on IoT messaging migrations, clustered MQTT infrastructure, and deployment automation. I also built monitoring tools that helped developers see how their systems behaved.
2018 — 2020
Intuit
DevOps Engineer · via third-party contractor
At Intuit, I built migration tooling that performed five times faster than the native solutions, along with regional registry monitoring and custom Artifactory extensions.
Customer-facing consulting, 2014 – 2017
2014 — 2017
Nth Generation Computing
Senior Solution Architect
I worked directly with enterprise customers to understand their operational needs, design solutions, and lead the implementations: VMware SRM disaster recovery with array-based replication, vRealize Operations, and Horizon VDI.
2014
VSS Monitoring
Senior Solutions Architect
I delivered a Nutanix-based virtual desktop platform for 300+ users with an expansion roadmap to 500+, and built the competitive positioning that explained the solution’s value to the customer.
How I like to work.
I like to start by talking with the people who use a system and understanding where they get stuck. From there, I work through the design, write the code, and stay involved as the team starts using it.
Before platform engineering, I spent years as a solution architect discovering what enterprise customers needed and owning the delivery. That thread runs through the AI tools I build today, and it is what draws me to forward deployed engineering.
Day to day that means Go, Python, and TypeScript on Kubernetes and AWS, with observability on VictoriaMetrics and Prometheus, and AI tooling built around MCP.
B.S. Computer Engineering New Jersey Institute of Technology
Let’s get in touch.
If my experience sounds useful to your team, I’d be happy to talk.
