Open to new roles / Remote

I build the platforms AI and multi-cloud products run on, and I’m having a blast.

Hi! I’m Shubham, a platform and site reliability engineer who has spent 7+ years making AWS, GCP and Azure behave. At Akka I led the redesign of our control plane on Crossplane, which now runs 40+ platform deployments for enterprise customers. I also own the platform side of our AI inference and fine-tuning product, and got to take it from a scrappy proof of concept all the way to production.

Hand me an ambiguous, high-stakes problem and a whiteboard, and I’m a happy engineer.

apiVersion: people.dev/v1
kind: Engineer
metadata:
  name: shubham-girdhar
spec:
  role: Senior Platform / SRE
  experience: 7y+
  clouds: [aws, gcp, azure]
  focus:
    - llm-serving-and-training-infra
    - multi-cloud-control-planes
    - kubernetes-operators
  languages: [go, python, scala, java]
status:
  phase: OpenToWork
  conditions:
    - type: Ready
      status: "True"
# CRD for kind Engineer, validates profile.yaml
apiVersion: apiextensions.k8s.io/v1
kind: CustomResourceDefinition
metadata:
  name: engineers.people.dev
spec:
  group: people.dev
  scope: Cluster
  names:
    kind: Engineer
    plural: engineers
  versions:
  - name: v1
    served: true
    storage: true
    subresources: { status: {} }
    schema:
      openAPIV3Schema:
        type: object
        properties:
          spec:
            type: object
            required: [role, experience]
            properties:
              role:
                type: string
              experience:
                type: string
                pattern: '^\d+y\+?$'
              clouds:
                type: array
                items:
                  type: string
                  enum: [aws, gcp, azure]
              focus:
                type: array
                items: { type: string }
              languages:
                type: array
                items: { type: string }
          status:
            type: object
            properties:
              phase:
                type: string
                enum: [OpenToWork, Running]
              conditions:
                type: array
                items:
                  type: object
                  required: [type, status]
                  properties:
                    type: { type: string }
                    status:
                      type: string
                      enum: ["True", "False", Unknown]
Years in platform and SRE
7+
Platform deployments at my current job
40+
Clouds, one control plane
3
Merged PRs to zio-http
36+

01 / Focus

What I love working on

AI / ML infrastructure

Turning GPUs into something product teams can just use. Multi-tenant LLM serving and fine-tuning as a platform feature: Go operators on Kubernetes DRA and vLLM, GPU pools that hop between hyperscalers and neoclouds, and an inference gateway on agentgateway that sees every model request.

  • vLLM
  • agentgateway
  • KubeRay
  • K8s DRA
  • CUDA / NCCL
  • LoRA

Multi-cloud platforms

I like making infrastructure the easy “yes” in an enterprise deal. Composite cross-cloud APIs on Crossplane that bake in each customer’s security, networking and compliance needs: hub-and-spoke networking, CMEK, zero-trust access.

  • Crossplane
  • Terraform
  • FluxCD
  • Cilium
  • Linkerd

Reliability and observability

Pagers should be quiet and dashboards should be cheap. Running production across dozens of regions with observability that pays for itself: self-hosted Cortex and Tempo, golden images, spot-node tricks, and OpenTelemetry end to end.

  • OpenTelemetry
  • Prometheus
  • Grafana
  • mTLS / PKI

02 / Experience

Places I’ve shipped things

  1. Apr 2023 - Present

    Remote

    Senior Site Reliability Engineer II at Akka (Lightbend)

    Akka Optimize (AI inference and training)

    • Own the platform side of Akka’s AI inference and fine-tuning offering, taking it from proof of concept to production.
    • Architected multi-tenant LLM serving as a native platform capability, managed end to end from the Akka CLI. Wrote the Go operators behind it (Kubernetes DRA, vLLM) and wired agentgateway’s external-processing hook into the gateway router.
    • Designed GPU capacity as a pool of independent clusters across AWS (H100, Crossplane-provisioned) and a neocloud (H200, no inbound ports); training runs land wherever there’s capacity.
    • Delivered the first enterprise design partner end to end, with models fine-tuned and served on the platform. That closed the product’s first platform deal.
    • Stay hands-on in product code across the trainer runtime (Python) and gateway router (Java, Akka SDK).

    Akka Platform

    • Led the redesign of the platform from Terraform to Crossplane, with composite cross-cloud APIs that unlocked sales to several large enterprise customers in finance, healthcare and aviation.
    • Operate the multi-cloud control plane behind 40+ platform deployments on AWS, Azure and GCP, including a 22-region footprint for a single customer spanning Asia, Canada and the US.
    • Own tenant onboarding: CMEK on GCP, Cilium networking, federation-plane permissions and customer-specific isolation.
    • Designed a cloud-agnostic zero-trust access gateway centralizing authentication, RBAC and audit logging.
    • Drove infrastructure and observability cost optimization while rolling out Linkerd and Groundcover across all customer regions.
  2. Jan 2026 - Jun 2026

    Remote (contract)

    Platform Engineering Consultant at PreSolv360

    • Built the AWS platform for a legal workflow management system from scratch: modular Terraform, EKS with IRSA, Aurora, Redis, SQS/SNS, GitOps with GitHub Actions and FluxCD, Kong with an OWASP WAF, Datadog.
    • Established the SOC 2 / ISO 27001 compliance baseline: CloudTrail, GuardDuty, AWS Config, Secrets Manager with External Secrets Operator.
  3. Oct 2021 - Mar 2023

    Mumbai

    Software Development Engineer 2 at Dream11

    • Led infrastructure engineering for d3.club, a play-to-earn cricket NFT game, from proof of concept to production, and built the NFT bridge exchanging token data with rario.com.
    • Wrote the entire infrastructure as code with Terraform CDK and orchestrated the game on Kubernetes; implemented region-level blocking with CloudFront Functions for US compliance.
    • Designed an internal developer platform to improve developer experience.
  4. Apr 2021 - Oct 2021

    Gurugram

    DevOps Engineer at Deutsche Telekom Digital Labs

    • Architected CI/CD on a Jenkins Shared Library with 80% test coverage.
    • Cut container image group size by 5 GB and in-house tooling LOC by 80%.
    • Improved alerting with Prometheus, Alertmanager and Grafana; ran Aerospike, Squid and Filebeat on Kubernetes; added Kafka monitoring via JMX.
  5. Jan 2019 - Apr 2021

    India

    Software Consultant, DevOps and Scala at Knoldus

    • Built a subscription-aware filtering microservice on Akka Streams and Akka Cluster over Kafka and Cassandra for an Internet-of-Vehicles platform.
    • Ran Spark at scale on auto-scaled EKS with shuffle-based autoscaling and pod-level RBAC.
    • Led real-time geospatial impression analytics on Spark, Kafka and Cassandra on Azure, and a Scala IoT controller on Lagom with CQRS and event sourcing.

References

Available upon request. Some endorsements can be found on my LinkedIn page.

04 / Toolbox

Tools I reach for

AI / ML infrastructure

  • vLLM
  • Ray / KubeRay
  • Kubernetes DRA
  • NVIDIA GPU stack
  • CUDA
  • NCCL
  • RDMA
  • agentgateway
  • Gateway API Inference Extension
  • LoRA serving

Platform and SRE

  • Kubernetes
  • Operators
  • Crossplane
  • Terraform
  • Ansible
  • FluxCD
  • Helm
  • Docker
  • GitHub Actions
  • Jenkins
  • Teleport
  • Kong
  • Linkerd
  • Cilium
  • AWS
  • GCP
  • Azure

Observability and security

  • OpenTelemetry
  • Prometheus
  • Grafana
  • Cortex
  • Tempo
  • Groundcover
  • Datadog
  • mTLS / PKI
  • CloudTrail
  • GuardDuty
  • AWS Config
  • WAF

Languages

  • Go
  • Python
  • Java
  • Scala
  • Akka
  • ZIO
  • Netty
  • Vert.x

Data and messaging

  • Kafka
  • Cassandra
  • PostgreSQL
  • Redis
  • MongoDB
  • Spark on Kubernetes
  • Elasticsearch
  • Aerospike
  • SQS / SNS

05 / Contact

Let’s build something

Hiring for platform, SRE or AI infrastructure? Wrestling with a gnarly multi-cloud or GPU problem? Or just want to geek out about Kubernetes? Drop me a note, I’d love to hear from you.

Rather talk it through? Grab a slot on my calendar.

Book a call