Forward Deployed
AI Engineer
Build. Deploy. Deliver.
Become the engineer who can take a real business problem and deliver a working AI solution.
7 weeks of intensive live engineering. 6 months of guided project implementation. Architecture, agents, enterprise integration, deployment and feedback — built for experienced software professionals.
Your Software Engineering Experience Is an Advantage.
Recommended for professionals with 8+ years of experience. Strong engineers with fewer years are welcome if they have solid programming and software engineering fundamentals.
Software Engineers
Frontend, backend, full-stack, platform and QA automation engineers.
Technical Leaders
Tech leads, staff engineers, architects and engineering managers who code.
Technical Problem Solvers
Technical PMs and consultants with significant hands-on coding experience.
Comfort with any established programming language is sufficient to start. Labs use Python; familiarity with APIs, Git and application development is expected.
AI Demos Are Easy. Delivery Is Engineering.
An FDE connects customer needs, software architecture and AI implementation — then makes the system work under real constraints.
One Program. Two Phases.
7-Week Live Engineering Intensive
28 instructor-led Zoom sessions, plus 10.5 hours of Saturday office hours.
- LLM applications and enterprise RAG
- LangChain, LangGraph and OpenAI Agents SDK
- Evaluation, security and production engineering
- Docker, CI/CD and AWS deployment
- Integrated capstone and architecture defense
6-Month Project Implementation Mentorship
Guided implementation clinics, project reviews and architectural feedback across six months.
- Three progressively challenging enterprise projects
- Customer-style requirements and constraints
- Implementation and architecture reviews
- Testing, evaluation and deployment improvements
- Portfolio demonstrations and technical defense
7 Weeks. 28 Live Sessions.
A focused engineering journey. Every week produces a working artifact, not just a set of notes.
WEEK 01 FDE Foundations & AI Solution Design
Move from receiving requirements to defining and designing AI solutions.
Session 1 — Forward Deployed Engineering
- FDE responsibilities and customer-facing engineering
- Problem discovery, stakeholder interviews and success metrics
- When AI is the wrong solution
Session 2 — LLM Engineering Essentials
- Models, tokens, context, inference and latency
- Prompt and context engineering
- Model selection, limitations and cost trade-offs
Session 3 — Building AI Services
- Python for experienced developers
- FastAPI and Pydantic
- OpenAI API, structured outputs and streaming
Session 4 — Solution Architecture Workshop
- Functional and nonfunctional requirements
- Architecture diagrams and design decisions
- Tool calling and service integration
WEEK 02 LangChain & LLM Application Engineering
Build reliable applications that connect models to real business functionality.
Session 5 — LangChain Fundamentals
- Models, prompts, messages and runnables
- Output parsing and structured generation
- Framework abstractions versus direct SDK usage
Session 6 — Tools & Function Calling
- Tool definitions, schemas and validation
- External APIs and business operations
- Timeouts, retries and error handling
Session 7 — Reliable LLM Workflows
- Context management and conversation state
- Model routing, fallback and provider independence
- Rate limits, token usage and cost control
Session 8 — Enterprise Integration Lab
- Connecting databases and REST APIs
- Authentication boundaries and secrets
- Testing tool-enabled applications
WEEK 03 Enterprise RAG & Knowledge Systems
Ground AI applications in enterprise knowledge instead of relying on model memory.
Session 9 — RAG Architecture
- RAG versus long-context prompting and fine-tuning
- Document ingestion, parsing and chunking
- Embeddings and similarity search
Session 10 — Vector Retrieval
- PostgreSQL and pgvector
- Metadata filtering and tenant isolation
- Semantic and hybrid retrieval patterns
Session 11 — Advanced Retrieval
- Query rewriting and reranking
- Context assembly and source citations
- Retrieval failure modes
Session 12 — RAG Evaluation
- Evaluation datasets and expected answers
- Retrieval relevance and groundedness
- Tracing and debugging with LangSmith
WEEK 04 Agentic AI, LangGraph & MCP
Design controlled AI workflows that can use tools, maintain state and request human approval.
Session 13 — Agent Architecture
- Workflows versus autonomous agents
- Agent loops, tool selection and execution boundaries
- When not to use an agent
Session 14 — LangGraph Engineering
- Graphs, nodes, edges and shared state
- Persistence, checkpoints and recovery
- Human-in-the-loop approvals
Session 15 — Model Context Protocol
- MCP clients, servers and tools
- Building and connecting an MCP server
- Authentication, permissions and tool safety
Session 16 — OpenAI Agents SDK
- Agents, tools, handoffs and guardrails
- Tracing and agent execution
- OpenAI Agents SDK versus LangGraph: architecture trade-offs
WEEK 05 Evaluation, Testing, Security & Reliability
Make AI behavior measurable, testable and safe enough for controlled deployment.
Session 17 — AI Evaluation Engineering
- Golden datasets and evaluation design
- Deterministic checks and LLM-as-judge limitations
- Quality, latency and cost metrics
Session 18 — LangSmith Evaluation & Tracing
- Execution traces and failure diagnosis
- Dataset-based experiments
- Regression testing for prompts, models and tools
Session 19 — AI Security
- Prompt injection and untrusted content
- Data leakage, access control and least privilege
- Tool misuse and approval controls
Session 20 — Reliability Engineering
- Timeouts, retries and circuit breakers
- Idempotency and safe workflow recovery
- Fallbacks, failure handling and incident scenarios
WEEK 06 Docker, GitHub Actions & AWS Production AI
Move from local execution to a deployed, observable AI service.
Session 21 — Containerization
- Dockerfiles, images and containers
- Environment configuration and secrets
- Local integration testing
Session 22 — CI/CD for AI Systems
- GitHub Actions workflows
- Automated tests and evaluation gates
- Build and deployment pipelines
Session 23 — AWS Deployment
- AWS ECS Fargate and container deployment
- IAM, Secrets Manager and service configuration
- Deployment verification and rollback
Session 24 — Production Observability
- CloudWatch logs and metrics
- LangSmith traces and AI quality monitoring
- Latency, throughput, token usage and cost
WEEK 07 FDE Capstone & Customer Delivery
Integrate the complete engineering stack and defend your solution as a customer-facing engineer.
Session 25 — Customer Requirements & Scope
- Ambiguous business requirements
- Acceptance criteria and success measures
- Scope, assumptions and delivery risks
Session 26 — Architecture & Implementation Review
- System integration and design review
- Model, retrieval and agent trade-offs
- Security and scalability decisions
Session 27 — Production Readiness
- Evaluation evidence and release criteria
- Monitoring, runbooks and failure recovery
- Documentation and handover
Session 28 — Capstone Demo & Defense
- Customer-facing solution demonstration
- Architecture decision defense
- Feedback and six-month implementation roadmap
Technologies You'll Work With
Focused hands-on engineering. Tools are chosen to solve problems, not to fill a checklist.
Models & APIs
OpenAI API, Python, FastAPI, Pydantic, structured outputs and tool calling.
Agents & Orchestration
LangChain, LangGraph, OpenAI Agents SDK and MCP.
Enterprise RAG
Embeddings, PostgreSQL, pgvector, retrieval and reranking patterns.
Evaluation & Testing
LangSmith, pytest, regression datasets, tracing and security testing.
Cloud & DevOps
Docker, GitHub Actions, AWS ECS Fargate, IAM, Secrets Manager and CloudWatch.
Explore Further
Hugging Face, Ollama, OpenRouter, Kubernetes/EKS and advanced AWS MLOps patterns where relevant to projects.
7 Weeks to Learn. 6 Months to Build.
Real engineering capability comes from implementation, feedback and iteration. The mentorship extends learning beyond the intensive.
Enterprise Knowledge AI
Design and implement an enterprise knowledge system with ingestion, retrieval, access controls, evaluations and operational tracing.
Focus: Reliable AI application engineering.
RAGpgvectorLangSmithIT Operations AI Agent
Build an agent that investigates incidents, interacts with operational tools and proposes actions with human approval.
Focus: Agent orchestration and reliability.
LangGraphMCPSecurityCustomer-Facing AI Solution
Respond to an ambiguous customer brief, handle changing requirements and deliver an evaluated, deployed solution.
Focus: End-to-end FDE delivery.
AWSArchitectureDeliveryHow Fortnightly Project Clinics Work
Sprint brief and acceptance criteria
Independent implementation
Live code and architecture review
Feedback, iteration and next milestone
Advanced topics such as local models, Hugging Face, Kubernetes and AWS MLOps are explored when justified by project requirements.
What You'll Have to Show
The goal is not a collection of course certificates. It's evidence of engineering capability.
Technical Portfolio
- Source-controlled AI applications
- Architecture diagrams and design decisions
- Automated tests and evaluation results
- Deployment and observability evidence
- Security considerations and operational documentation
Forward Deployed Skills
- Translate business problems into technical requirements
- Choose and defend architectural trade-offs
- Integrate AI into existing software systems
- Present working solutions to stakeholders
- Respond to feedback and changing constraints
Live Schedule & What's Included
7-Week Intensive
12 November – 25 December 2026
Monday, Tuesday, Thursday & Friday
9:00 PM – 10:30 PM IST
Saturday Office Hours
9:00 PM – 10:30 PM IST
Classes and office hours are conducted live on Zoom. Saturday office hours are optional.
The final session falls on 25 December; any holiday-related adjustment will be communicated to enrolled learners.
Program Includes
- 42 hours of live instructor-led training
- 10.5 hours of Saturday office hours
- 6 months of fortnightly project mentorship
- Hands-on labs and weekly assignments
- Project, code and architecture reviews
- Session recordings and source code
- Capstone guidance and portfolio development
Learn with Sanjay Negi
Enterprise AI Architect · Software Engineering Leader
With 22+ years in enterprise software engineering, architecture and production systems, Sanjay teaches AI through the lens of engineering decisions, real implementations and operational trade-offs.
The focus is simple: understand the system, build it, test it, deploy it and explain why it works.
Frequently Asked Questions
Clear answers before you commit.
Is 8+ years of experience mandatory?
Do I need prior AI or ML experience?
What programming language will be used?
Will we learn LangChain, LangGraph and OpenAI Agents SDK?
Will Docker, AWS and CI/CD be covered?
Will we cover Hugging Face, Ollama and OpenRouter?
What happens during the six-month mentorship?
How much time should I commit?
Will I become a Forward Deployed Engineer after this course?
What if I miss a live class?
Are additional cloud or model API charges included?
Where can I see the fee?
Stop Collecting AI Tutorials.
Start Delivering AI Systems.
Bring your software engineering experience. Build the AI engineering and customer-delivery capability to take on more complex problems.
Claim 75% Diwali Discount ↗7 weeks live training · 6 months project mentorship · Online via Zoom
Refund Policy
This program is offered at a discounted price. All enrollments are final and non-refundable. Please review the curriculum, schedule and program commitments before registering.