00/ 05
Bridge
Preparation
What you do
- Production Python, Git, REST with FastAPI, SQL, Docker
- LLM fundamentals on local models
- Screening test. Below the bar gets a bridge plan, not a seat.
Modules & labs
- Python for Engineers
- SQL, APIs & Git
For engineering teams and sponsoring leaders
Forward Deployed AI Engineering
Scope and ship AI inside a real client, with a deployed system the client retains.
For engineers with 2+ years of experience. B2B teams and limited individual seats.
Programme details
- Duration
- 24 weeks · 250 hours
- Delivery
- Live online 3 times a week
- Cohort 1
- Tuesday 12 January 2027
- Credential
- SUNY Potsdam Professional Certificate
This programme is not HRD Corp claimable.
02Who is an FDE?
What is a Forward Deployed Engineer?
A Forward Deployed Engineer is a software engineer who works inside a customer's organisation to scope, build, ship and run AI systems against their real data and workflows, and owns the outcome in production.
- 01 – Build
Software engineering
Build production-ready systems.
- 02 – Deploy
AI and ML engineering
Make them reliable, measured and usable.
- 03 – Solve
Consulting
Adapt them to a real client's problem.
Build•Deploy•SolveForward Deployed Engineer
03Curriculum overview
What does the programme cover?
The Forward Deployed Engineer Course in six facts
- What it is
- Train, then deploy
- Certificate
- SUNY Potsdam
- Who it is for
- Engineers, 2+ years
- Format
- 24 weeks, live online
- Delivery cadence
- Live online 3 times a week
- Next cohort
- Tuesday 12 January 2027
01/ 05
The engagement method
Engagement method
What you do
- Palantir-style decomposition of vague problems
- Discovery workshops and operating-model maps
- The eval set written with the client before any build
- PRD, statement of work and pricing a pilot
Modules & labs
- The Forward Deployed Engineer Mindset
- Solution Documentation & Communication
02/ 05
Production LLM systems
Production LLM systems
What you do
- Context engineering and structured outputs
- RAG over scanned PDFs, tables and drawings
- Evaluation suites and regression gates
- AI-assisted engineering with Claude Code and Cursor
- TypeScript and React for engineers
Modules & labs
- Production Software Engineering for FDEs
- Large Language Model Fundamentals
- Embeddings, Vector Search & RAG
- Evaluating LLM Applications
- AI-Assisted Engineering
- Build and test a FastAPI service with a Streamlit front end that wraps a public dataset — deployed locally with Docker.
- Enterprise document Q&A assistant with hybrid RAG, citations, and an evaluation dashboard.
- GraphRAG over a product/customer knowledge graph (Neo4j).
03/ 05
Agents, data and integration
Agents, data and integration
What you do
- LangGraph agent loops with human-in-the-loop
- MCP servers over internal systems
- Data engineering: SQL, Spark, dbt
- SSO, ServiceNow, Salesforce and SAP integration
Modules & labs
- Tool Calling & Function Calling
- Agent Architectures & Orchestration
- Enterprise Systems Integration
- Data Engineering for AI Solutions
- Customer-support agent with MCP tools (order lookup, ticket creation, refund policy retrieval), human approval gates, and guardrails.
- Multi-agent research & report-generation workflow with LangGraph.
04/ 05
Deploy, secure, operate
Module 4
What you do
- AWS in depth; private, on-prem and sovereign patterns
- Cost SLAs, observability and CI/CD
- Prompt-injection defence, PII and the DPDP Act
- A graded incident drill
Modules & labs
- Guardrails, Safety & Reliability for Agents
- Cloud Deployment & DevOps for AI
- LLMOps: Observability, Cost & Performance
- Fine-Tuning, Open-Weight Models & Private Deployments
- Containerise and deploy the customer-support agent to AWS with CI/CD, tracing, caching, and a cost dashboard.
- Fine-tune and serve a small open-weight model for a domain classification task; compare against a prompted frontier model on cost and accuracy.
05/ 05
Forward deployment
Forward deployment
What you do
- Your pod of 4–5 is embedded with a real client
- Discovery, walking skeleton, eval gate
- Harden, go live, hand off
- Graded by the client sponsor
Modules & labs
- Domain Solution Patterns
- FDE Interview Preparation
- Career Track & Placement Support
05Technologies
Tools and technologies covered in the FDE course
54tools in 10 families, wired the way a client system is built
Languages & frameworks7
PPythonFFastAPIPPydanticSStreamlit+3
LLM APIs & models6
OOpenAIACAnthropic ClaudeGGGoogle GeminiLLlama+2
Agent & orchestration6
LLangGraphLLangChainOAOpenAI Agents SDKAAAnthropic Agent SDK+2
Retrieval & data7
PPineconeQQdrantWWeaviatePpgvector+3
Evaluation & observability6
RRAGASDDeepEvalLLangSmithLLangfuse+2
Cloud & DevOps6
DDockerAAWSAOAzure OpenAIVAVertex AI+2
Fine-tuning & serving4
HFHugging FaceLQLoRA/QLoRAVvLLMOOllama
Integration6
O2OAuth 2.0SOSAML/OIDCKSKafka/SQSSHSalesforce & HubSpot APIs+2
AI-assisted development3
CCClaude CodeCCursorGCGitHub Copilot
Data engineering3
APAirflow/PrefectDdbtSWSQL warehouses
07The FDE way
The foundation of every deployment
Which real engagements will you study?
Each week opens with a real AiSPRY engagement, taught by the engineer who ran it. Then you build it yourself.
- Rela HospitalPharma bot on RAGCase study
- CognigenAgentic hospital administrationCase study
- Dhammika PereraInsurance and quotation-SLA agentsCase study
- GMR EnergyDay-ahead demand and price forecastingCase study
- Uttarakhand GovernmentStatewide GenAICase study
- Siemens and ReitzDocument and vision AI on the shop floorWeeks 3 and 7
Live, not recorded
250 hours: 144 live, 60 in labs, and a one-month client engagement.
LiveLabsClient
Client craft, practised weekly
- Client Craft Lab role-plays
- Interview Gym · 6 mocks
- Vertical elective · BFSI, health, support, public sector
- Anthropic Claude Developer exam prep
0
API bills to pay
Every exercise runs on free local models. A shared frontier-API credit covers comparisons and the client engagement.
11Outcomes
What you leave with
Proof, not a transcript: a system running at a client, three graded artifacts, and eight capabilities you can demonstrate.
- 250hours
- live, with labs
- 3artifacts
- production-grade, graded
- 1live system
- deployed at a real client
- 8capabilities
- you can prove at interview
Where it leads
- You todayEngineer, 2+ years
- Programme completionLive system · SUNY Potsdam certificate
- Roles this prepares you for
Forward Deployed EngineerApplied AI EngineerAI Solutions EngineerCustomer EngineerDeployed EngineerAgent Engineer
01 · Discovery
Run a structured discovery with a business stakeholder and convert it into a scoped, prioritised AI solution plan.
02 · LLM systems
Build production-grade LLM applications using RAG, structured outputs, and tool calling.
03 · Agents
Design and ship agentic AI systems (single-agent and multi-agent) with guardrails, memory, and human-in-the-loop controls.
04 · Integration
Integrate AI solutions with enterprise systems — REST/GraphQL APIs, databases, ERPs/CRMs, SSO, and event streams.
05 · Deployment
Containerise, deploy, secure, and monitor AI services on cloud infrastructure with CI/CD.
06 · Evaluation
Build evaluation and observability pipelines that prove business value and detect drift, regressions, and cost blow-ups.
07 · Client craft
Present demos, write solution documents, handle objections, and drive adoption inside a client organisation.
08 · Interviews
Clear FDE interviews at AI-first companies and enterprise consulting firms.
12Certificate
A credential you earn, not one you attend.
What certificate do you get, and how is it earned?
State University of New York at PotsdamIssued with 360DigiTMG
Credential
Professional Certificate in Forward Deployed AI Engineering
Verification
QR code on every certificate and a public register at 360digitmg.com/verify
- Requirement 01
≥ 70%Client deployment on the published rubric
- Requirement 02
≥ 60% eachThree production artifacts
- Requirement 03
≥ 60%Proctored exam, 60 questions
- Requirement 04
PassClient Craft Lab role-plays
Earn the certificate by passing all four assessments.This professional certificate is not a degree or visa pathway.
Rubric and exam blueprint (PDF)Plan your team's participation
Discuss your team's experience, client setting and deployment goals with us.
Talk to us about your teamRelated programme: Certified AI Governance Professional (CAIGP).
What stood out most was the programme's practical and structured approach to AI adoption. The industry-specific examples, combined with clear frameworks for evaluating feasibility and designing solutions, made the learning highly relevant to our business.
What stood out most was the trainer's simple way of explaining Data Science and GenAI, and showing us different AI tools that I had not used before. It made me realise how useful these tools can be in our daily work.