Job description
About Zinnov
Zinnov is a global management consulting firm that helps organizations make decisions that actually get used — and deliver results that matter. For over two decades, we’ve partnered with leading enterprises, high-growth technology companies, and investors to answer some of the toughest questions they face:
Where should we invest?
How do we scale globally?
What capabilities will win in the next decade?
Our work shapes market entry strategies, global operating models, M&A decisions, and long-term growth bets. We’re known for being data-led, execution-focused, and outcome-driven** — not opinion-heavy slideware.
At Zinnov, how you work matters as much as what you deliver. We value independent thinking, crisp communication, and early ownership. With 450+ professionals across 10 global offices, we work across industries including Digital Services, ER&D, Enterprise Software, Semiconductors, Healthcare, BFSI, Automotive, Media & Telecom, and Private Equity.
Zinnov isn’t for everyone. It’s for people who want steep learning curves, honest feedback, and the chance to see their work influence real business decisions — not just presentations.
** About the Role **
As an ** Agent / AI Engineer ** , you will build the AI layer of the GCC Intelligence Platform—the conversational agent users interact with. You will own grounding, retrieval, persona routing, session state, and permission-aware querying so the agent stays accurate, secure, and tenant-safe.
** What You’ll Do **
** LLM & Agent Integration **
- Integrate LLM APIs (e.g., AzureOpenAIor equivalent) for conversational workflows. • Build a grounding layer that anchors responses to real platform data (not model guesses). • Maintain prompt templates across multiple personas and use-cases.
** Retrieval, Permissions & Security Boundaries **
- Implement intent classification and persona routing to the right KPI views. • Build the API handler that sets DB sessionvariables so RLS/ABAC policies enforce correctly. • Own JWT validation and permission registry lookups that control access.
** Reliability & Cost Controls **
- Build session state management, safe retries, and failure handling. • Implement token budgeting, tracking, and enforcement per tenant. • Improve accuracy baselines through structured evaluation and regression checks.
** What You Bring **
** Qualifications & Experience **
- 3+ years building with LLM APIs in production (OpenAI/Azure/Anthropic or similar). • Experience shipping grounding/RAG systems to real users. • Strong Python for agent logic, APIs, and prompt management. • Understanding of how auth/session context interacts with database queries.
** Key Skills **
- Prompt engineering across user types and tasks. • Accuracy-first mindset—grounded correctness as a constraint, not a feature. • Strong debugging and evaluation discipline.
** What Success Looks Like – Global Excellence (GE) **
- Agent answers staygrounded—accuracyimproves and hallucinations reduce. • Access is safe—tenant boundaries are respected every time. • Persona routing feels natural—users get the right view, right detail, right format. • Costs stay controlled—token governance and caching/optimization reduce overrun risk.
If you enjoy building secure, grounded AI agents that executives can trust, and shipping AI systems with strong evaluation discipline — this role offers both impact and progression.
Zinnov is an equal opportunity employer.** We celebrate diversity and are committed to building an inclusive workplace. We welcome applications from individuals of all backgrounds, communities, and experiences.