Companies in this catalog do not charge applicants. Never send money to apply.
Title: Senior Agentic Engineer
Location: Remote (UTC+0 - UTC+3)
Brahma AI operates at the intersection of enterprise Media Asset Management (MAM) and cutting-edge generative media. We build and scale industry-leading generative AI models, including hyper-realistic digital humans (ATMAN) and multilingual voice synthesis (VAANI), for world-class enterprise clients in entertainment, sports, healthcare, and retail.
Role Overview
We are looking for a Senior Agentic Engineer to build, ship, and run the components of our internal agentic setup: gateway and routing features, tool connectors, and governed agent workflows that let engineers across Brahma use AI agents safely. Much of this work happens in our sovereign agent gateway for model routing, MCP tool hosting, and multi-agent orchestration.
This is a senior individual contributor role. You will build to designs set by the Architect for Agentic AI, own well-scoped components end to end from prototype to production, and contribute to design reviews and architecture decision records (ADRs). You keep up with new agent tooling and patterns, trial the ones that hold up, and make them production-ready for enterprise use. You also share what you learn, within the team and with other Brahma teams.
As a senior individual contributor, you will own components end-to-end from initial prototype to production. Working to technical designs set by our Architect for Agentic AI, you will contribute actively to architecture decision records (ADRs) and design reviews, vet emerging agent frameworks, and harden the most effective patterns for enterprise scale.You won’t have to carry the entire stack alone. You’ll be embedded in a dedicated, high-caliber unit:
Lead Engineer for Agentic AI: Team lead driving operational delivery and SLOs.
Architect for Agentic AI: Sets high-level architecture and strategic technical direction.
Product Management: Directs priorities and feature roadmaps.
ML Engineers: Own underlying models, evaluation frameworks, and context techniques.
Security Engineer for Agentic AI: Defines governance standards and security controls.
Infrastructure & DevOps: Maintains the foundational cloud platforms while you manage agent runtimes.
We assemble our setup from open protocols, open-source components, and our own services, so we stay in control of security, cost, and vendor choice. You build what is missing and maintain what you run, including upgrades, patches, and incident response.
Key Responsibilities
Platform & Integration Engineering
Gateway & Routing Features: Build and maintain features such as provider fallbacks, budget enforcement, usage attribution, and access policies.
MCP Servers & Connectors: Build and operate MCP servers and connectors to Git, CI/CD, work tracking, and documentation, so agent output arrives as reviewable, traceable change.
Deployment & Runtime: Own CI/CD, containers, and infrastructure-as-code for agentic AI services, on the cloud platform run by Infrastructure & DevOps.
Agent Workflows & Context
Governed Workflows & Multi-Agent Patterns: Implement agent workflows with structured inputs and outputs and human approval gates, and multi-agent patterns (A2A, supervisor/worker) with loop, budget, and failure controls.
Context Engineering & Memory: Build and run memory and context services (session and long-term memory, context compaction, retrieval, progressive tool disclosure) to the architecture set by the Architect, using techniques evaluated by the ML Engineers.
Production Ownership: Run the components you ship after launch: monitoring, alerting, rollbacks, cost, and incident response, and keep them current with upgrades.
Quality, Observability & Security
Tests & Evals in CI: Write unit and integration tests, and regression evals for agents, tools, and routing using the evaluation framework owned by the ML Engineers, and run them in CI.
Observability Pipeline: Own the tracing and usage-reporting pipeline across model, tool, and agent calls, so the Security Engineer can detect threats, ML Engineers can monitor quality, and the Lead can track SLOs and spend.
Secure by Default: Build to the security standards and controls owned by the Security Engineer for Agentic AI, and fix security findings in the components you own.
Operations & Adoption
Maintain What You Run: Keep services and open-source dependencies patched, diagnose failures to root cause, and take part in incident response and any future on-call rotation.
Developer Surfaces & Catalog: Build and maintain the ways engineers use the setup: the CLI, chat integrations, coding-agent harness configuration, and the shared catalogue of agents and skills.
Coding Agent Adoption: Help teams across Brahma adopt coding agents through documentation, working examples, and pairing.
Knowledge Sharing: Document what you build, and share patterns and lessons within the team and with other Brahma teams through demos, guides and pairing.
Must Haves
Experience: 5+ years of software engineering experience, with primary expertise in server-side Python development.
LLMs in Production: Hands-on experience shipping production features built on LLM APIs and tool calling, including their failure modes and cost.
MCP & Connectors: Experience building MCP servers or equivalent tool connectors, not only consuming them.
APIs & Async Systems: Proven track record building reliable APIs, services, and async or event-driven systems.
Testing & Observability: A disciplined approach to automated testing, CI, logging, metrics, and tracing, and to debugging production issues.
Security Fundamentals: Practical knowledge of OAuth 2.0 and OIDC, secrets management, least privilege, and LLM-specific risks such as prompt injection.
Knowledge Sharing: Track record of documenting your work and helping colleagues and other teams adopt new technology.
AI Coding Literacy: Pragmatic daily use of AI coding tools (e.g., Claude Code, Cursor, Copilot) while keeping code quality, test coverage, and review standards high.
Multi-Agent & Workflow Engines: Experience with A2A, agent frameworks (e.g., LangGraph, Google ADK), or durable workflow engines (e.g., Temporal, Prefect).
Memory & Retrieval: Experience with agent memory designs (session, long-term, compaction), vector databases (e.g., pgvector, Qdrant), or hybrid search.
Coding Agent Harnesses: Experience configuring coding agents for teams: skills, subagents, hooks, and sandboxing.
Nice to Have
Gateways & Routing: Experience running a multi-model gateway or routing layer in production.
Evaluation & Tracing Tooling: Experience with LLM evaluation tools, LLM-judged regression tests, and OpenTelemetry.
Industry Trends: Track record of bringing emerging agent tools or patterns from evaluation into production, with a clear view of when not to adopt them.
Cloud & Containers: Experience running services on GCP (AWS or Azure transferable) with Docker, Kubernetes, and Terraform or equivalent IaC.
