UTKARSH SAXENA

My work connects product decisions with infrastructure engineering: turning operational requirements into Go services, containerized workloads and reliable distributed systems.

Munich, Germany (Remote). I lead the development and roadmap execution of PioWorkers, a digital workers and cloud operations platform. Alongside product decisions, requirements and prioritization, I contribute to backend architecture and software development.

Piovation engineering / contributions

Piovation / private repository contributions

GPU allocation and resource lifecycles. My infrastructure work includes GPU allocation, Kubernetes workloads and cloud resource management. The important boundaries are the resource request, the placement decision, the running workload and the point at which capacity can be released again.

The walkthroughs are simplified logical views based on the supplied work history and infrastructure context. Recovery and capacity scenarios illustrate behavior; they are not production incident reports.

A workload declares its resource requirements. The scheduler checks available GPU capacity before placing it.

RequestAllocateCapacityRecoverRelease

Logical architecture / select a scenario to explorePause

Cloud worker orchestration. I designed and developed Golang microservices for infrastructure management, worker orchestration, scheduling and operational automation. The control plane coordinates work while containerized workers execute it, with lifecycle and progress information feeding back into the operation.

A cloud operation enters the control plane, which separates orchestration from execution and schedules the work.

ScheduleDispatchObserveRetryComplete

Logical architecture / select a scenario to explorePause

Container infrastructure and deployment reliability. I architected and optimized Kubernetes deployments, including pod lifecycle operations, resource allocation and cluster reliability. I also built Docker-based infrastructure to improve deployment consistency, scalability and operational efficiency across environments.

A service change is packaged into a consistent Docker image. The versioned artifact becomes the input to deployment.

BuildRoll outReadyUnhealthyRecover

Logical architecture / select a scenario to explorePause

Product ownership and engineering practices. I coordinate cross-functional delivery, mentor engineers, conduct architecture and code reviews, and establish engineering practices that improve code quality and delivery. Technical problem-solving spans distributed systems, performance optimization and cloud resource management.

Kolkata, India. I worked on independently deployable Golang services, asynchronous pipelines and edge applications, with REST and RabbitMQ connecting the services and Linux-based unit testing supporting maintainability.

Concurrent microservices and event processing. I optimized blocking operations with goroutines and channels, with a reported 40% throughput improvement. Event-driven pipelines combined RabbitMQ, worker pools and concurrent consumers for fault tolerance under load.

Independently deployable Go services accept REST operations and publish asynchronous work through RabbitMQ.

PublishProcessRetryCacheRespond

Logical architecture / select a scenario to explorePause

Cloudflare and edge applications. I collaborated on Cloudflare Vector AI training in a Workers / Node.js environment and integrated it with WordPress CMS, with a reported 30% reduction in bounce rate. I also built the SEO-focused Dofollo web application with React and TypeScript, Cloudflare Workers, D1 for edge storage, and a supporting MongoDB microservice. The resume reports a 10% reduction in hosting costs.

Service communication. I implemented Protocol Buffers and gRPC serialization to reduce peak-usage transaction latency by 30 ms. This work focused on efficient communication between services alongside the concurrency improvements in the processing layer.

Remote, India. I developed Golang microservices for RabbitLoader with RabbitMQ for asynchronous communication and Firebase Authentication integration, contributing to a 100-page speed score.

High-traffic cloud services. I built and managed services on DigitalOcean and Cloudflare supporting an average of 60,000 daily interactions.

Database and caching performance. I used MongoDB, MySQL, Redis, BigCache and Memcache to achieve sub-second response times. The resume reports a 40% reduction in query latency and a 25% improvement in system throughput for this work.

Remote, India. I developed Node.js and Express.js backend services using Firebase Pub/Sub on Google Cloud Platform for multiplayer gaming with more than 100 simultaneous users.

APIs and cloud service integration. I designed REST APIs with half-second response times and automated bulk testing in Postman to verify reliable communication between the cloud services.