Engineer
Run AI in production.
Control cost and latency, orchestrate models, roll out changes safely and test for regressions.
Build AI · Tier 3 · Bengaluru residency
AI Expert. Pod-ready in 90 days.
Rooman's AI Expert is a 360-hour, 3-month in-person residency in Bengaluru and Tier 3 of the Build AI track. You learn production AI engineering, agentic systems, RAG, AI-augmented engineering, client-side delivery and AI security, then spend your capstone inside a real Rooman product team and earn the Rooman AI Engineer Certificate.
360 hours7 modulesIn person, BengaluruCapstone project
What you'll be able to do
Become pod-ready.
AI Expert is the bridge between AI Professional and a real client pod. Over 90 days in person in Bengaluru, you work on production agentic systems, retrieval, AI-assisted engineering, client delivery and AI security, and your capstone is four weeks inside a Rooman product engineering team.
Build
Design agents and retrieval that hold up.
Build agents with LangGraph, CrewAI and the OpenAI and Claude Agent SDKs, and RAG systems you can measure.
Deliver
Work like a pod engineer.
Write specs and design docs, work async with international clients, and ship features inside a Rooman product team.
Skills covered
- Production agentic systems with failure recovery
- RAG: chunking, hybrid retrieval, reranking and evaluation
- AI-augmented engineering with Cursor and Claude Code
- Client-side delivery and code review
- AI security and the OWASP LLM Top 10
- Production AI deployment and observability
Who it's for
Built for engineers going pod-ready.
The course is for engineers who have completed AI Professional, Tier 2 of Build AI, and want to work in a client-facing AI engineering pod. Admission is selection-based. You need strong English for client work and the commitment to a 90-day in-person residency in Bengaluru, and an engineering degree is strongly preferred.
Roles where these skills are used
- Senior AI Engineer (pod-ready)
- Production ML Engineer
- AI Solutions Architect
- Rooman GCC Pod Engineer
- StartupVarsity Founding Engineer
- AI Product Engineer
Entry requirements.
- Completion of AI Professional (Tier 2)
- A selection interview and portfolio review
- Strong English communication for client-side delivery
- Commitment to a 90-day in-person residency in Bengaluru
New to AI? Start with AI for ALL, Rooman's 20-hour, 9-module foundation course. Not sure which course fits? Take the 10-minute career assessment.
360 hours · 7 modules
A curriculum built from production AI to a real product team.
Seven modules, all in person in Bengaluru, ending in a capstone inside a Rooman product team.
01Production AI engineering at scale60 hrs
Keep AI systems fast, affordable and safe to change in production.
Topics covered
- Cost engineering: token budgets, prompt caching and semantic caching
- Multi-model routing and fallback chains; latency and streaming SLOs
- Gradual rollout with feature flags, canary deployments and A/B tests
- Regression testing with golden datasets, eval-on-commit and production logging
You produce: a prototype hardened for 10,000 users a day within a monthly budget.
02Agentic systems deep build70 hrs
Design agents that plan, use tools, recover from failure and hand off to people.
Topics covered
- Agent patterns: ReAct, Plan-and-Solve, ReWOO and Tree-of-Thought
- LangGraph for state-machine agents and CrewAI for multi-agent systems
- OpenAI Agents SDK, Claude Agent SDK and MCP server authoring
- Failure recovery, human-in-the-loop and long-horizon agents with memory
You produce: a customer-support agent that handles 5 task types within a cost per ticket.
03RAG and retrieval mastery50 hrs
Diagnose why RAG fails, and build retrieval you can measure.
Topics covered
- Chunking strategies: fixed, semantic, hierarchical and late chunking
- Embedding models and vector databases: pgvector, Pinecone, Qdrant, Weaviate
- Hybrid retrieval (BM25 + dense), reranking and query rewriting
- Retrieval evaluation: precision@k, MRR, NDCG and Ragas
You produce: a 100,000-document RAG system that meets a precision target.
04AI-augmented engineering at senior pace40 hrs
Use AI coding tools the way senior engineers do, and know when not to.
Topics covered
- Spec-driven development: writing specs an AI can execute
- Cursor and Claude Code in depth, and reviewing AI-written code
- Test-first engineering, and planner, coder and reviewer agent workflows
- When not to use AI tools: security-critical paths and novel design
You produce: a feature shipped faster with AI tools, with no more defects than hand-written code.
05Client-side delivery in international pods40 hrs
Work with US, EU, UAE and Singapore clients the way a pod engineer does.
Topics covered
- Client expectations and cultural calibration in international pods
- Writing PRs, design docs and RFCs at senior level
- Async work, client repositories and sprint planning across time zones
- Live demos, client presentations and business English
You produce: a week shadowing a real Rooman GCC pod.
06AI security, governance and regulated industries30 hrs
Defend AI systems and meet the rules that apply to them.
Topics covered
- OWASP LLM Top 10 and prompt-injection defence
- PII detection, redaction and audit logs in LLM workflows
- GDPR, HIPAA, SOC 2 Type II and the EU AI Act for AI systems
- Adversarial robustness, red-teaming and abuse monitoring
You produce: a red-team report on a Rooman product AI feature.
07Capstone in a real Rooman product pod70 hrs
Four weeks inside a Rooman product team, from orientation to a shipped feature.
Topics covered
- Pod assignment: HireAI, BlueLinked, AI Tutor or CRM Voice Agent
- Two sprints: paired feature work, then independent delivery with code review
- Final demo to the product engineering lead and a pod-ready assessment
- Placement match: GCC pod, Founders' Track or hiring partner
You produce: a shipped feature, presented to the product engineering lead.
Tools covered
The tools you'll work with.
Every tool is used hands-on in the residency labs and your capstone.
| Tool group | Tools | What you use them for |
|---|---|---|
| Agent frameworks | LangGraph, CrewAI, OpenAI Agents SDK, Claude Agent SDK | Build single-agent and multi-agent systems |
| Protocols | MCP (Model Context Protocol) | Connect agents to tools and data, with governance |
| Retrieval | pgvector, Pinecone, Qdrant, Weaviate, Ragas | Store, search and evaluate retrieval |
| Serving and MLOps | FastAPI, Triton, MLflow, Kubernetes | Serve, track and scale models |
| AI coding tools | Cursor, Claude Code | Write, test and review code at senior pace |
Capstone
A real product. A real team.
Spend four weeks embedded in a Rooman product engineering team, on HireAI, BlueLinked, AI Tutor or CRM Voice Agent. You deliver features across two sprints, demo your work to the product engineering lead and receive a pod-ready assessment.
Every module also ends in its own capstone or lab, from hardening a prototype to red-teaming an AI feature.
Your capstone portfolio.
- A production-hardened AI prototype
- A customer-support agent across 5 task types
- A 100,000-document RAG system
- A red-team report on an AI feature
- Your capstone: a feature shipped in a Rooman product team
Your credential
Rooman AI Engineer Certificate.
With your capstone portfolio and a recommendation. The product engineering lead you worked with adds a recommendation to your certificate and portfolio.
Your final assessment says whether you are pod-ready. If you need more time, the residency extends by one supervised four-week sprint.
What's included
Everything you need to finish.
- Certified industry-expert trainers
- In-person residency on Rooman's Bengaluru campus, about 6 hours a day
- In-person mentorship and fully equipped labs
- AI-powered LMS with 1-year access
- A capstone inside a Rooman product team
- Placement matching: GCC pod, Founders' Track or hiring partner
Before you enrol
Questions pod-ready engineers ask.
Can't find your answer? Call 080 6945 1000 or WhatsApp us.
Who is this course for?
The course is for engineers who have completed AI Professional, Tier 2 of Build AI, and want to work in a client-facing AI engineering pod. Admission is selection-based. You need strong English for client work and the commitment to a 90-day in-person residency in Bengaluru, and an engineering degree is strongly preferred.
Is AI Expert only in person?
Yes. All seven modules run in person on Rooman's Bengaluru campus, because the capstone places you inside a real Rooman product team.
What is the selection process?
An application and portfolio review, a 45-minute technical interview and a 30-minute behavioural interview.
Can I take AI Expert without AI Professional?
No. AI Professional is a hard prerequisite, because you need production engineering depth before going pod-ready.
How much time does it take?
360 hours over 3 months (90 days), about 6 hours a day in person.
What is the capstone?
Spend four weeks embedded in a Rooman product engineering team, on HireAI, BlueLinked, AI Tutor or CRM Voice Agent. You deliver features across two sprints, demo your work to the product engineering lead and receive a pod-ready assessment.
What if I don't pass the final assessment?
The residency extends by one supervised four-week sprint to close the gaps.
Where can I go after AI Expert?
The final module includes a placement match to one of three pathways: a Rooman GCC pod serving international clients, the StartupVarsity Founders' Track, or Rooman's hiring partners.
Does the course include placement support?
Yes: placement matching to a Rooman GCC pod, the StartupVarsity Founders' Track or a hiring partner. It is not one of Rooman's job-guaranteed programmes.
Which certificate will I receive?
The Rooman AI Engineer Certificate, with your capstone portfolio and a recommendation from the product engineering lead.
What is the integrated programme?
AI Foundation, AI Professional and AI Expert can be taken together as an integrated six-month programme, which includes a stipend during the AI Expert capstone. Admission to it is selection-based, and a counsellor can explain the options.
How do I find out fees and batch dates?
Talk to a Rooman counsellor on 080 6945 1000 or WhatsApp +91 97390 86029. They will share current fees, payment options and upcoming batch dates.
Related courses.
All Build AI tiers
120 hrsTier 1AI FoundationBuild AI Tier 1: the next step into AI engineering, in 120 hours.
240 hrsTier 2AI ProfessionalBuild AI Tier 2: one of five specialisation tracks, in 240 hours.
20 hrsStart hereAI for ALLThe 20-hour foundation for every Apply AI specialisation.For companiesNeed an AI engineering pod?Top Build AI graduates join Rooman's GCC pods. Talk to us about an AI engineering pod for your company.Talk to usNext step
Start with a conversation.
Tell us your degree or current role, and a counsellor will call you with batch dates, fees and payment options.
- Call: 080 6945 1000
- WhatsApp: +91 97390 86029
- Email: hello@rooman.com
Course details reviewed on 25 September 2026.