Skip to content
BB
Ben BakhtierPlatform Engineering

Infrastructurethat makes thesafe path fast.

I build the internal platforms product teams ship on. Kubernetes, AWS, delivery systems, and the AI agents that now run on top of them.

View career journeyLinkedIn ↗
CKA, CKAD + 2x AWS + Terraform5 verified certificationsPlatform engineeringKubernetes, AWS, Terraform, ArgoCDBuilt on AWSTerraform, CloudFront + S3, Next.js, Docker
0+
Years building platforms
AWS · Kubernetes · Terraform
0+
Engineers served
Backstage golden paths
12d → <1d
Service provisioning
Self-service Terraform modules
0+
Repos on AI review
Claude Code + GitHub Actions
25-35%
Kubernetes compute saved
Karpenter + Spot for non-prod
01

About

I build the systems that let engineers ship faster and operate with confidence.

With 6+ years of hands-on experience I design and operate cloud-native infrastructure on AWS and Kubernetes. On an 8-person platform team, I build the golden paths, Infrastructure as Code, and GitOps delivery that let 250+ product engineers ship without filing infrastructure tickets.

Today I'm extending that platform model with AI agents, applying them to code review, infrastructure operations, and incident diagnosis, with security, scoped permissions, and human approval built into the execution path.

I'm focused on building the next generation of platform engineering: infrastructure that is easier for engineers to use, and AI that can safely participate in operating it.

Operating principles
01

The safe path is the fast path

Guardrails should not be speed bumps. Golden paths win when they are genuinely the shortest route to production.

02

Self-service over ticket queues

Recurring infrastructure work is a product bug. Turn it into a module, a template, or a workflow engineers can run themselves.

03

Agents need scoped permissions

AI in operations is only viable with RBAC, plan review, and human approval on the execution path. Returning "unknown" beats guessing.

02

Career journey

Six years from supporting regulated healthcare workloads to building the platform, and the agents, that 250+ engineers ship on.

  1. Sep 2023 to Present
    CleverTap
    Customer engagement and analytics SaaS platform
    Mountain View, CA
    3 years

    1. Senior Platform Engineer

      Sep 2025 to Present

    2. Cloud Engineer II, Platform Infrastructure

      Sep 2023 to Aug 2025

    One of an 8-person platform team building the golden paths, Infrastructure as Code and GitOps delivery that let 250+ product engineers ship without filing infrastructure tickets, plus the AI agents that now run on top of it.

    • Built Backstage golden paths and reusable Terraform modules used by 250+ engineers, cutting new-service provisioning from 12 days to under 1.
    Claude CodeMCPBackstageArgoCDHelmEKSKarpenterTerraformIPv6OIDCSpot
  2. Oct 2019 to Aug 2023
    Komodo Health
    Healthcare data and analytics platform
    New York, NY
    3 years 11 months

    1. DevOps Engineer, Cloud Infrastructure

      Jun 2021 to Aug 2023

    2. Junior DevOps Engineer

      Oct 2019 to May 2021

    Cloud and Kubernetes infrastructure for healthcare data and analytics in a HIPAA-regulated environment, from first Linux and Terraform foundations through to owning production.

    • Ran AWS infrastructure for healthcare data analytics across development, staging, and production, with least-privilege IAM, KMS encryption, and Secrets Manager under HIPAA.
    HIPAAIAMKMSSecrets ManagerGitLab CIJenkinsDockerHelmPrometheusGrafana
Full resume
03

Capabilities

What I build with, and what I have run in production.

Platform & Developer Experience

Backstage golden paths, self-service provisioning, reusable Terraform modules, and internal tooling that removes infrastructure from the critical path.

BackstageGolden pathsTerraform modulesSelf-service workflows

Kubernetes & Delivery

EKS at production scale with GitOps delivery: ArgoCD and Helm across 50+ microservices, replacing scheduled releases with repeatable deployments.

EKSArgoCDHelmGitOpsRBAC

Cloud Infrastructure as Code

AWS designed and operated entirely through Terraform: private IPv6 networking, OIDC-authenticated pipelines, and PR-based change control.

AWSTerraformIPv6OIDCCloudFormation → TF

AI in Infrastructure

Agents applied to code review, operations, and incident diagnosis, with scoped MCP servers, RBAC, plan review, and human approval on every mutating path.

Claude CodeMCP serversScoped RBACHuman-in-the-loop

Observability & Resilience

EFK logging with parallel S3 archival that survives Elasticsearch failures, plus Prometheus and Grafana coverage built for the on-call engineer.

PrometheusGrafanaEFKS3 archival

Security & Compliance

Least-privilege IAM, KMS encryption, and Secrets Manager protecting sensitive healthcare data in HIPAA-regulated production environments.

HIPAALeast-privilege IAMKMSSecrets Manager
04

Certifications

Every credential below links to its public verification page.

CKAVerify ↗

Certified Kubernetes Administrator

The Linux Foundation

Aug 2026expires Jul 2028
ID 9865a321-e14a-4d8b-afd8-6e7f044ca651
CKADVerify ↗

Certified Kubernetes Application Developer

The Linux Foundation

May 2026expires May 2028
ID 32d674eb-96ba-4d23-9460-1dedf35a2b60
AWS SAAVerify ↗

AWS Certified Solutions Architect - Associate

Amazon Web Services

Aug 2026expires Aug 2029
ID e6ad55be-2f43-4359-96a4-de0b18834311
Terraform AssociateVerify ↗

HashiCorp Certified: Terraform Associate (004)

HashiCorp

Feb 2026expires Feb 2028
ID 03564919-c050-42cf-af02-c27cac20d37f
AWS CCPVerify ↗

AWS Certified Cloud Practitioner

Amazon Web Services

Mar 2026expires Mar 2029
ID 2e9c9b65-1c5d-446d-929b-814c1074b1e3
05

Selected work

Case studies on platform tooling, cost engineering, and putting AI agents into workflows without giving up control of them.

  1. 70+
    repositories, up from 30

    AI Pull-Request Reviewer

    An agent that reviews every pull request against its Jira acceptance criteria, and says Unknown when it cannot tell.

  2. RBAC
    on every mutating path

    Namespace-Scoped MCP Servers

    Letting AI agents operate Kubernetes without letting them operate it unsupervised.

  3. 12d → <1d
    to provision a new service

    Backstage Golden Paths

    Cutting new-service provisioning from twelve days to under one, by making the safe path the fast one.

  4. 25-35%
    Kubernetes compute saved

    Karpenter and Spot Capacity

    Cutting Kubernetes compute spend by 25-35% without moving critical services off On-Demand.

All case studies
06

Digital twin

Ask about my experience and get an answer grounded in my actual career history, not a generic chatbot.

Online

I'm Ben's digital twin. I answer from his real experience: the roles, the numbers, the systems he actually built. Ask me anything about his work.

Enter to send · Shift+Enter for a new line

Answers are generated by openai/gpt-oss-120b via OpenRouter, constrained to Ben's documented history. It will say when something isn't covered rather than guess, but for anything that matters, verify with Ben directly.

07

Let's build the platform.

Open to senior platform and DevOps engineering conversations, especially teams thinking seriously about developer experience and about AI that can safely participate in operations.

bakhtier.ben@gmail.comin/benbakhtier ↗