Skip to content
Search jobs inside your AI agent. Claude, ChatGPT, Claude Code, Codex, Gemini, Cursor and Grok
Connect

QuanXAI

NewBe an early applicant

DevOps and SRE (Thai speaking)

Posted

Description

Job Title: Site Reliability Engineer

(SRE / DevOps / Backend — AWS Infrastructure & Client Deployments)

QuanXAI | Bangkok, Thailand | Full-time, Hybrid (partial work from home)


About Us

About USWe are an AI Tech-Startup providing AI and ML product and consulting services. As an AWS Selected Consulting Partner, we deliver tailored consultancy and end-to-end solutions that align with each customer’s unique business requirements, leveraging best-in-class cloud technologies.Our solutions are designed and deployed on native cloud or hybrid architectures, built around the AWS Well-Architected Framework, focusing on security, cost optimization, scalability, reliability, and operational excellence. Beyond AI/ML, we also deliver strategic data initiatives, including data analytics, data warehousing, and data integration from ERP and SCM systems.We are 100% cloud-focused, with deep expertise in integrating data from on-premises systems. Our certified AWS professionals bring over 20 years of experience in designing, implementing, and delivering enterprise projects across the APAC region.


Job Responsibilities

As an SRE (DevOps / Backend), you will take on the following:


1. AWS Infrastructure & Cloud Platform (30%)

• Multi-Account Architecture: Design and operate our AWS Organizations landing zone — account-

per-environment isolation (dev/staging/prod/per-client), Service Control Policies, IAM Identity

Center (SSO), and centralized logging/audit (CloudTrail, Config) — so that a mistake in one account

can never take down another.

• Infrastructure as Code: Own our Terraform estate end-to-end — reusable modules, remote state

with locking, per-environment workspaces, drift detection in CI, and policy-as-code gates

(Checkov/OPA). Console clicking is for emergencies only, and every emergency becomes code

afterward.

• Networking & Connectivity: Design VPC topologies, Transit Gateway, PrivateLink, and Route 53;

build and operate secure connectivity into customer sites — Site-to-Site VPN, Direct Connect, and

zero-trust tunnels (Tailscale/SSM) — respecting IT/OT segmentation and the Purdue model on plant

networks.

• Compute & Orchestration: Run production EKS — Karpenter/cluster autoscaling, node pools

including GPU nodes for model inference, pod security standards, resource quotas — plus

ECS/Fargate and Lambda where they fit better. Right tool, not favorite tool.

• Security & Compliance: Enforce least-privilege IAM, KMS encryption everywhere, Secrets Manager

with rotation, GuardDuty/Security Hub findings triage, and WAF — and keep our posture aligned

with client security assessments (ISO 27001-style ISMS controls, PDPA).

• FinOps: Own the AWS bill — tagging discipline, CUR/Cost Explorer dashboards, right-sizing, Savings

Plans and Spot strategy, storage lifecycle policies — because at a startup, burn rate is an SLO too.


2. Release Engineering & Client-Site Image Versioning (30%)

• Container Image Lifecycle: Own the full image supply chain in ECR — multi-stage, multi-arch builds

(amd64/arm64), immutable tags with semantic versioning, and a promotion pipeline (dev → staging

→ prod → client release channels) where an image is built once and promoted by digest, never

rebuilt.

• Supply Chain Security: Sign every release image (cosign), generate SBOMs (Syft), scan continuously

(Trivy/ECR scanning), and run a base-image patching cadence — so when a client's security team

asks “what exactly is running in our plant?”, the answer is one command away.

• Customer Fleet Version Management: Maintain the version matrix of which client runs which

release of which component — across cloud tenants, on-prem servers, and edge boxes. Design per-

client configuration overlays (Helm/Kustomize) with GitOps (ArgoCD/Flux), staged rollouts by

release ring, pinned versions for conservative clients, and one-command rollback with database

migration safety (expand/contract, backward-compatible APIs).

• On-Prem & Air-Gapped Delivery: Deliver and update our platform inside customer environments

with no or restricted internet — registry mirroring, offline image bundles, K3s on industrial PCs,

store-and-forward telemetry — and schedule update windows around plant shifts and production

runs, because “just redeploy” doesn't exist on a factory floor.

• Release Discipline: Run the release train — changelogs and release notes clients can actually read,

versioned APIs and schema migrations, canary/blue-green deployments in cloud, and a

compatibility policy (N-1/N-2 support) so field versions never drift beyond what we can support.


3. SRE Practice & Advanced DevOps (25%)

• SLOs & Error Budgets: Define SLIs/SLOs for our platform (availability, latency, inference success

rate, data pipeline freshness) per environment — including client sites — and use error budgets to

arbitrate between shipping features and paying down reliability debt.

• Observability: Build the single pane of glass — Prometheus/Grafana metrics, Loki/CloudWatch logs,

OpenTelemetry tracing across services and LLM calls — with remote fleet health visibility from

every customer site that allows it, and alerts tuned so a page always means something.

• Incident Management: Own on-call (we keep it humane and compensated), write runbooks before

incidents and update them after, lead blameless postmortems, and drive MTTR down with better

automation — not more heroics.

• CI/CD at Scale: Evolve GitHub Actions pipelines — caching, matrix builds, ephemeral preview

environments per PR, automated integration/load tests (k6) — and track DORA metrics

(deployment frequency, lead time, change failure rate, MTTR) as the team's scoreboard.

• Resilience Engineering: Design and test for failure — backup/restore with measured RTO/RPO,

multi-AZ architecture, chaos experiments and game days, and documented disaster-recovery

procedures for both cloud and on-prem deployments.


4. Backend Engineering (15%)

• Platform Services: Design and build the internal services that power operations — deployment

orchestrator APIs, license/activation services for client installs, and telemetry ingestion from

customer sites — in Python (FastAPI) or Go, with proper auth (OAuth2/JWT, mTLS between

services).

• Data Layer Operations: Operate PostgreSQL in production — HA/failover, PITR backups, zero-

downtime migrations, query/index tuning — plus Redis and time-series databases (TimescaleDB)

used by our industrial workloads.

• Performance Engineering: Profile and optimize hot paths — connection pooling, caching strategy,

async I/O, queue-based decoupling (SQS/Redis Streams/Kafka) — so the platform stays fast as client

count grows.


Background / Experiences

• Engineering Foundation: B.Eng. or B.Sc. in Computer Engineering, Computer Science, or a related

field — with strong fundamentals in operating systems, networking (TCP/IP, DNS, TLS), and

distributed systems, plus deep comfort in Linux.

• Production Ownership: You have run real systems in production — carried on-call, debugged

incidents at 2 AM, and improved the system so that page never fires again. You measure yourself by

uptime and lead time, not lines of code.

• IaC & Kubernetes in Anger: Hands-on Terraform and Kubernetes experience beyond tutorials —

you have designed modules, untangled state, debugged CrashLoopBackOffs, and know why

resource limits and PodDisruptionBudgets exist.

• Shipping to Customer Environments: Experience delivering versioned software to environments

you don't fully control — on-prem, edge, air-gapped, or heavily regulated — is a major advantage.

You respect that someone else's production is sacred.

• Startup Temperament: Comfortable with ambiguity, broad scope, and pragmatic trade-offs; you

communicate clearly with technical and non-technical colleagues in Thai and English, and document

as you go.


Knowledge & Skills

• AWS Depth: EKS/ECS, EC2, VPC/Transit Gateway, IAM, RDS, S3, Lambda, CloudWatch,

Organizations/Control Tower, KMS/Secrets Manager — with the judgment to know when managed

beats self-hosted.

• Containers & GitOps: Docker internals (layers, caching, multi-stage), Kubernetes (networking,

storage, RBAC, Helm/Kustomize), and GitOps workflows with ArgoCD or Flux.

• IaC & Automation: Terraform (modules, state, workspaces), CI/CD with GitHub Actions or GitLab CI,

and scripting in Python, Go, and Bash.

• Observability Stack: Prometheus/Grafana, Loki or ELK, OpenTelemetry, and alerting design

(Alertmanager/PagerDuty) — including SLO-based alerting.

• Security Fundamentals: TLS/PKI, secrets management, image scanning and signing, network

segmentation, and least-privilege thinking applied by default.

Bonus Points (Preferred if you have)

• AWS Certifications: Solutions Architect Professional, DevOps Engineer Professional, or Security

Specialty.

• Industrial / OT Experience: Deployments on plant networks — OT/IT segmentation, MQTT brokers,

OPC UA connectivity, K3s on industrial PCs, or time-series databases (TimescaleDB, InfluxDB).

• GPU & Model Serving Ops: GPU node management (NVIDIA device plugin, MIG/time-slicing), model

servers (vLLM, Triton), and inference load testing.

• Advanced Tooling: HashiCorp Vault, service mesh (Istio/Linkerd), Crossplane, Nix, or progressive

delivery tooling (Argo Rollouts, Flagger).

• Compliance Exposure: Supporting ISO 27001 ISMS controls, vendor security assessments, or

PDPA/GDPR data-handling requirements.

• Craft Signals: Open-source contributions, a serious homelab, incident write-ups or technical blog

posts demonstrating systems thinking.


QuanXAI — Making enterprise AI adoption seamless, secure, and measurable.

Click Apply or send resume to Nopawan@Quanxai.com

Skills

Apply

See also

DevOps jobs by country — openings, pay and top skills →