Principal AI Systems Engineer
About Us: Founded in 1972, Atari is one of the world’s most iconic consumer brands and a pioneer in the video game industry, known for creating classics like Pong, Asteroids, and Centipede. Today, Atari Inc. continues to build on its legacy by developing games, hardware, and experiences that honor the past while driving innovation for the future.
Over the past two years, we've been building Atari India, a growing team that plays a critical role in supporting our global operations. We're proud of the team we've assembled so far, and we’re just getting started. As part of a lean, high-impact organization, the team in India works closely with colleagues in North America and Europe on projects that move the company forward. Whether you're helping launch a new game, keeping our infrastructure secure, or supporting day-to-day operations, your work here matters. Join us as we continue to grow Atari India and build the future of a legendary brand.
Position: Principal AI Systems Engineer
Experience: 8+ Years
Location: Netaji Subhash Place, Pitampura, New Delhi
Employment Type: Full-Time (Hybrid)
Reports to: Senior Director of Technology, India
About the Role
Architect, build, and own AI systems that automate expert-intensive technical workflows end-to end — from CLI frameworks, MCP servers, and agent tooling through to production deployment, business outcome tracking, and continuous improvement. You solve real business problems with AI, ensure solutions are fully implemented and adopted, and measure whether they are actually working.
Responsibilities
System Architecture
- Own end-to-end architecture of AI automation systems: workflow decomposition, component communication, human checkpoints, and failure behaviour
- Design and build internal CLI frameworks, reusable libraries, and agent scaffolding
- Author and maintain agent instruction files (SKILL.md, CLAUDE.md, system prompts) and MCP server definitions
- Configure Claude Code and Codex CLI environments: MCP wiring, tool permissions, slash commands, and engineering standards
- Evaluate and document architectural trade-offs across reliability, latency, cost, and maintainability
Pipeline Development
- Build production-grade AI pipelines in Python: orchestration, structured prompting, context assembly, schema validation, and retry strategies
- Integrate AI systems with external tooling — version control, build pipelines, SDKs, compliance databases, internal APIs
- Design context assembly: how domain knowledge, runtime state, retrieved documents, and tool outputs compose into the precise input each pipeline stage needs
- Build and operate multi-agent systems: orchestrator-worker patterns, agent memory, structured handoffs, and conflict resolution
Prompt & Context Engineering
- Design, version, and maintain system prompts and agent instructions as first-class engineering artefacts
- Own output schema design and prompt regression testing with a maintained ground-truth eval set
- Engineer context windows with precision — balancing accuracy, token cost, and latency through compression and selective retrieval
- Partner with the RAG Engineer to define retrieval requirements — what knowledge is needed, under what conditions, and at what granularity
- Build and maintain structured runtime knowledge assets: curated document corpora, rule sets, decision trees, and validation reference libraries
- Work with domain experts to translate specialist knowledge into agent behaviour: decision logic, edge cases, and failure modes
Evaluation & Reliability
- Build and own the evaluation framework: test suites, regression benchmarks, LLM-as-judge pipelines, and per-stage quality metrics
- Implement production monitoring using LangFuse, Arize, or equivalent — latency, token usage, success rates, and output quality drift
- Run structured failure analysis and implement targeted fixes across context assembly, orchestration, and tool integration
- Define automation rate as a first-class metric and report on business effectiveness of deployed systems
Governance & Technical Leadership
- Implement full audit trails — inputs, tools called, outputs, and human review triggers
- Enforce versioning of all agent instructions and system prompts as engineered artefacts with controlled rollout
- Set the technical standard for AI development across the organization — architecture patterns, eval practices, and quality gates
- Collaborate with engineering, product, and domain teams; engage leadership on roadmap priorities and technical risk
Requirements
- Proven track record of building production AI automation systems from scratch — end-to-end from architecture through deployment
- Hands-on expertise with Claude Code, Codex CLI, Cursor, or equivalent — including MCP server configuration and agent instruction authoring
- Experience designing and deploying MCP servers and custom tools: tool schema, authentication, and permission boundaries
- Experience building internal CLI frameworks, agent scaffolding, and reusable libraries that others build on
- Experience creating internal tooling and automation that measurably improved engineering team efficiency — reducing manual processes and accelerating workflows
- Experience working with data scientists and domain experts to implement AI solutions that measurably improved team productivity
- Deep prompt and context engineering: system prompts, few-shot design, chain-of-thought, token budget management, and prompt versioning
- Proficiency with LLM orchestration frameworks — LangChain, LangGraph, LlamaIndex, AutoGen, or equivalent
- Experience building AI evaluation frameworks: test suites, regression benchmarks, LLM-as-judge, and production quality monitoring
- Production Python engineering: modular, testable, well-logged code with proper error handling
- Cloud platform experience (AWS, Azure, or GCP): deploying and monitoring AI workloads with containerisation
- Experience integrating AI systems with external APIs — tool definition, permission management, and failure handling
- Experience defining and tracking AI productivity metrics: automation rate, time-to-completion, and human intervention rate
Bonus Points
- Experience in gaming: game development pipelines, Unity/Unreal engine architectures, or platform certification processes
- Familiarity with game engine scripting, asset pipelines, or platform SDKs (Xbox GDK, PlayStation SDK, or similar)