Asutosh Panda - Resume

Skills

GPU & AI Infrastructure: NVIDIA GPU Operator, DCGM, drivers/CUDA, MIG, vLLM, KubeRay/Ray Serve, LiteLLM, Vertex AI, GPU SLOs
Infrastructure & Networking: Kubernetes (GKE, EKS, bare-metal), Helm, Kustomize, FluxCD, ArgoCD, Terraform/OpenTofu, Ansible, Docker/Compose, Traefik, AWS ALB/NLB, Linkerd, GCP, AWS, Cloudflare, Hetzner, OVH
Secrets and Auth: Workload Identity Federation, IRSA + EKS Pod Identity, External Secrets Operator (ESO), GCP + AWS Secret Manager, cert-manager, Pomerium, Cloudflare Zero Trust (Tunnel + Access)
CI/CD and Observability: GitHub Actions, self-hosted ARC runners, Trivy, SonarQube, TruffleHog, Renovate; Prometheus, Grafana, Mimir, Loki, Tempo, Grafana Alloy, grafana-operator, Alertmanager
Development & Data: Golang, Python, FastAPI, Bash, Node.js, TypeScript, PostgreSQL, AWS Aurora, Redis, GCP Pub/Sub, Linux

Professional Experience

One2N Pune, Maharashtra, India

Senior SRE & Team Lead · Client: Fravity AI (Remote) May 2026 - Present

Stood up the model-inference platform on the org's new AWS EKS GPU cluster via Flux GitOps: KubeRay for vLLM serving, an HA LiteLLM gateway on Aurora + Redis (secrets via ESO), and a DCGM + Alloy observability spoke to the central LGTM.
Built fravity-spine, the production multi-tenant GKE platform for an AI product: private control plane, CMEK etcd, Workload Identity Federation, OpenTofu across 6 state layers, FluxCD add-ons, Pomerium SSO, and identity-based kubectl (Connect Gateway) that retired bastions and shared SSH.
Led multi-cloud governance and cost visibility across 5 GCP + 5 AWS accounts with the 2-engineer team I lead: audited posture (19 service-account keys, ~$26,520/yr zombie spend), shipped an org-wide FinOps board (per-account cloud, SaaS + AI spend in Grafana), migrated all human + 10 workload identities to Terraform IAM-as-code, and rebuilt CI on keyless Workload Identity Federation.
Delivered the central LGTM observability platform: Mimir + Loki + Tempo (GCS-backed, multi-tenant), Grafana unified alerting to Slack, and dashboards-as-code. A single grafana/alloy agent unifies the GKE spine/spokes and the first VM into one label-differentiated view (host/container/log).
Stood up a 5-tool SAST stack on self-hosted ARC runners and shipped scan-aggregator (normalises 7 scanners to a per-repo audit); delivered a customer security review surfacing 76 H/C findings across 4 images.
Delivered staging as a commit-to-URL GitOps pipeline (Flux image-automation, live TLS, Slack alerts), keyless via per-app Workload Identity; onboarded the full app tier including the workflow-engine with attribute-filtered PubSub isolation, and locked the database to automation-only writes via Atlas migrations.
Shipped the first client MVP on a hardened, fully IaC GCE VM (no public IP, IAP-only SSH, least-priv SA) behind a Cloudflare tunnel, with a keyless WIF build-to-deploy Ansible pipeline, the org's first reusable VM-deploy blueprint in an all-Kubernetes estate.

Luxor Technology Seattle, WA, US

DevOps Engineer (Remote) June 2023 - April 2026

Operated a multi-cluster Kubernetes platform for 20+ teams and 3,000+ enterprise customers: 8+ clusters across Hetzner, OVH, and GCP with 24/7 uptime for L1/L2 validator and RPC workloads, including 2 Hetzner bare-metal HA clusters (KubeOne, OpenEBS) built from scratch. Standardised on a 7-layer FluxCD GitOps pipeline with Linkerd mTLS, migrating 70+ secrets to Workload Identity.
Owned the on-call and incident platform: migrated 5 teams from Opsgenie to BetterStack, led response through real outages (a Litecoin hard fork with a zero-downtime deploy, a kubelet TLS expiry), and shipped x509 monitoring across 10 clusters.
Owned CI/CD for 50+ repositories on self-hosted ARC runners. Ported a failing Playwright E2E suite from Mac hosted runners ($5,000/mo) to Linux ARC, cutting runtime 40 to 15 min and cost to $600/mo. Built code-signed Windows + Mac release pipelines and automated FluxCD deployments to GCP Artifact Registry.
Built org-wide platform tooling: KubeOrbit (Go CLI for kubeconfig provisioning across 20+ clusters) and Memphis (Python SOC1/2/SOX compliance binary, solo in 4 weeks for the IPO). Vendored Trivy into a composite action across 9 repos, sidestepping the later CVE-2026-33634 in aquasecurity/trivy-action.
Proposed and built BVT, the org's first zero-hardcoded-secrets cluster: replaced SOPS-in-git with ESO + GCP Secret Manager + Workload Identity Federation; the pattern became the gold standard for new clusters. Built custom Python tooling against the Hetzner + OVH firewall APIs for secure cross-cloud bare-metal peering where no Terraform modules existed.

Tata Consultancy Services Hyderabad, Telangana, India

Systems Engineer (Remote) October 2020 - June 2023

Migrated 2,500+ repositories (GitLab, Bitbucket, SVN) to GitHub Enterprise for 120 teams and built their CI/CD pipelines, resolving self-hosted-API pagination limits across the platforms and saving 15 weeks of manual effort.
Built PowerShell + Python automation (server monitoring, alerting, log rotation, Oracle DB replication, MS SQL updates) across 22 RHEL/WebLogic servers and 3 projects, saving 380+ man-hours/month.
Shipped a Golang CLI that cut HP storage provisioning from 30 man-hours across 30 monthly tickets to a single command, adopted by the storage team as their primary provisioning interface.

Projects

Frit (May 2026 - Present) - GPU SRE lab: vLLM on NVIDIA GPU Operator; OSS PRs to dcgm-exporter + gpu-operator.
Platform Zero (Apr 2026 - Present) - 4 SRE labs (k8slab, ansiblelab, awslab, gcplab); 0-to-100 in one command.
Truss (Jun 2026 - Present) - Open-source self-hostable Supabase/Appwrite alternative, with an AI-agent MCP server.
ducat (Jun 2026 - Present) - Open-source Prometheus cost exporter: cloud, SaaS + AI spend in one Grafana view (FOCUS spec).

Teaching Experience

Full-Stack Instructor, 4 cohorts (2022-2026), taught NextJS, FastAPI, and DevOps to BSc CS students.

Education

SASTRA Deemed University Master of Computer Applications (MCA) - Chennai, India - 2021 - 2023