Senior AI DevOps / Platform Engineer (AI Agents & MCP)

Convo; Islamabad Capital Territory; On-site; PKR 3,500,000 - 6,000,000; Full time

Benefits offered by this job

Medical benefits

Free lunch

Annual increments

Awards

Agile culture

Diversity

Job summary

Convo is seeking a senior platform engineer to build, scale, and maintain production delivery platforms for AI Agents, LLM workloads, and MCP servers. You will deploy across Kubernetes, design CI/CD pipelines, implement GitOps, and enforce release gates with automated checks.

You will also instrument observability and security, manage secrets and sandboxing, and coach peer teams on release standards, aiming for high availability and zero-downtime rollouts.

Qualifications

• Minimum 9+ years in DevOps/SRE/platform engineering.

Responsibilities

• Deploy, host, scale, and operationalize AI Agent frameworks and MCP servers.
• Design CI/CD pipelines for Python services, agents, prompts, tools, and MCP configurations.
• Implement GitOps deployments with promotion, rollback and environment-specific policy controls.
• Integrate automated unit, API, security and LLM-evaluation checks into release gates.
• Implement secrets handling and sandbox isolation for agent execution.
• Instrument pipelines and services for observability, health monitoring, and incident diagnosis.
• Maintain reusable deployment templates and coach teams on release standards.

Skills

Kubernetes

GitOps

Infrastructure as Code

CI/CD

Python scripting

Terraform

Observability

Security

Cloud operations

Release engineering

Tools

LangFuse

Phoenix/Arize

OpenTelemetry

Datadog

Prometheus

Azure

Job description

Technical mission

Build, scale, and maintain production delivery platforms and infrastructure for AI Agents, LLM workflows, and Model Context Protocol (MCP) servers. Establish repeatable evaluation gates, enforce sandboxed agent execution, enable zero-downtime rollouts, and ensure high-availability runtime operations for non-deterministic AI workloads.

Key responsibilities

• Deploy, host, scale, and operationalize AI Agent frameworks, vector stores, and Model Context Protocol (MCP) servers across Kubernetes clusters.

• Design CI/CD pipelines for Python services, agent definitions, prompts, tools, MCP servers, and platform configuration.

• Implement GitOps-based Kubernetes deployments with promotion, rollback and environment-specific policy controls.

• Integrate automated unit, API, security and LLM-evaluation checks (e.g., prompt stability, tool-use accuracy, output safety, latency/cost regression) into release gates.

• Implement secrets handling, dynamic sandbox isolation for agent execution, and least-privilege deployment identities.

• Instrument pipelines and deployed services for agent-level observability (token usage, tool invocation tracking, LLM latency, cost monitoring, distributed tracing), health monitoring, and incident diagnosis.

• Maintain reusable deployment templates for services, agents, and MCP tools, and coach engineering teams on release standards.

Required technical capabilities

• Minimum experience: 9+ years in DevOps, SRE or platform engineering, including 3+ years with production Kubernetes, GitOps, and infrastructure as code, combined with hands-on production experience deploying AI Agents, LLM infrastructure, or MCP servers.

• Hands-on experience deploying, managing, and scaling AI-driven solutions, vector databases, proxy gateways, or MCP servers in production.

• Strong CI/CD and GitOps experience for Kubernetes-hosted services.

• Terraform or equivalent infrastructure-as-code and practical cloud operations experience, preferably Azure.

• Python scripting for pipeline tooling, release checks, operational automation, and integrating AI evaluation SDKs.

• Secrets management, container security, workload-isolation practices, and dynamic agent sandboxing.

• Monitoring, alerting, and telemetry for both delivery pipelines and runtime services (including LLM/Agent tracing tools like LangFuse, Phoenix/Arize, OpenTelemetry, Datadog, or Prometheus).

• Experience making automated test and LLM evaluation results enforceable in CI.

Preferred experience

• Release strategies such as canary, blue/green and progressive delivery for agent prompt changes and LLM versioning.

• LLM or agent evaluation frameworks (e.g., DeepEval, Ragas, LangSmith) and benchmark execution in CI.

• Software supply-chain controls, artifact signing and dependency/security scanning.

Expected deliverables / acceptance evidence

• Reusable CI/CD templates for services, agents, MCP tools, and framework releases.

• Automated quality, security and LLM-evaluation gates.

• GitOps deployment definitions with tested rollback paths.

• Release observability, AI tracing, audit trail and operational documentation.

Primary interfaces

Works with Infrastructure, Agent Runtime, QA Automation, AI Test & Dataset Engineering and service owners across all pods.

What We Have For You

• Great compensation package, medical benefits for you and your family, free lunch, annual performance-tied increments, performance recognition awards, and a great lean and agile work culture!

• Convo endorses a culture of diversity in all aspects and aims to build a diverse team of amazing individuals!