Job Responsibilities
Build AI solutions and Agents end to end: graphs, tools, MCP servers, and the services behind them, from design through production and iteration.
Practice the working craft of agents and their harness daily: context management, prompt design, plan construction, tool result shaping, and guardrails.
Build and maintain the evaluation sets and regression suites that measure whether an agent is improving; investigate failures down to the step that caused them.
Work directly with domain experts: collect the questions practitioners really ask, map them to the right capability, and resolve ambiguity at the source.
Debug from traces: reproduce, isolate, fix, and add the test that prevents recurrence.
Automate the development loop itself: turn repeated manual work into commands, hooks, and subagents.
Move decisions forward with working prototypes.
Requirements
A track record of building and shipping agentic systems that reached real users.
Deep practical craft with models: context management, prompting, plan generation, structured outputs, and tool design.
Strong Python; productive in an unfamiliar codebase quickly.
Experience running agentic workloads on a major cloud platform; ours can be learned quickly.
Fluent daily use of agentic coding tools; Claude Code preferred.
Clear written communication under time pressure.
Demonstrated self-direction: work they initiated, built, or fixed without being asked.
Preferred
Production experience across multiple model families.
Working TypeScript and solid SQL.
Front end capability: React, TypeScript, CopilotKit or similar.
Data engineering on large transactional datasets.
Evaluation and observability tooling.
Contributions to agent frameworks or MCP servers.
Tech Stack & Tools
Orchestration: LangGraph, Google ADK, Microsoft Agent Framework
Models: Anthropic, Google, OpenAI, open-weight families
Cloud: GCP Vertex AI, Azure AI Foundry
Protocols: MCP, A2A, AG-UI, A2UI
Engineering: Python, TypeScript, SQL, Claude Code
Delivery: evaluation harnesses as CI, tracing and observability tooling
Circana Behaviors
Beyond technical skills, experience, and role-specific attributes, these shared behaviors are fundamental to our culture and success. We seek individuals who consistently demonstrate and champion these behaviors in their daily work:
Stay Curious: Being hungry to learn and grow, always asking the big questions.
Seek Clarity: Embracing complexity to create clarity and inspire action.
Own the Outcome: Being accountable for decisions and taking ownership of our choices.
Center on the Client: Relentlessly adding value for our customers.
Be a Challenger: Never complacent, always striving for continuous improvement.
Champion Inclusivity: Fostering trust in relationships engaging with empathy, respect, and integrity.
Commit to each other: Contributing to making Circana a great place to work for everyone.
Location: This position can be located in the following area(s): Bangalore
Prospective candidates may be asked to consent to background checks (in accordance with local legislation and our candidate privacy notice ) Your current employer will not be contacted without your permission.