Assignments use generic profiles such as one, two, four, or eight-plus 80GB data-center GPUs. Students justify requests by model size, VRAM, concurrency, storage, latency, and whether the task is training, inference, RAG, or voice/agent hosting.
Noncredit · CDCP · enhanced funding
// Stackable pathway — noncredit CDCP Certificate of Completion with credit CTE Certificate of Achievement planning track
Build the infrastructure behind production-grade AI systems.
AI Infrastructure and Architecture is a 10-course, 540-contact-hour sequence that takes learners from foundational computing through inference serving, model adaptation, agentic systems, RAG, AI security, and shared AI Foundry operations.
Compute & containers
Cloud & MLOps
Inference serving
AI Foundry
STEM research studio
Agents & RAG
AI security & governance
No personal paid hardware required. Foundational labs use open tools and free tiers; advanced projects can use institution-managed shared compute.
10courses
540 hrscontact hours
30-unit equiv.credit equivalent
3 semestersstackable pathway
01 / overview
What the certificate covers
A 16-week-per-course sequence from foundational computing to production AI systems.
The AI Infrastructure and Architecture certificate is a 10-course, 540-contact-hour noncredit CDCP sequence with a planned credit CTE Certificate of Achievement pathway pending local curriculum approval. Students build the compute, container, cloud, data, serving, adaptation, agentic, retrieval, security, and governance skills that underlie modern AI workloads. Each 16-week, 54-hour course pairs lecture with hands-on labs using free and open-source tools, centrally managed cloud or AI Foundry resources, and no required personal paid software, cloud, GPU, or API accounts. Every course produces direct evidence mapped to program outcomes, culminating in a final employer-readable capstone portfolio.
02 / ai foundry
An AI Foundry turns classroom labs into shared production infrastructure.
The curriculum remains platform-agnostic, but advanced courses are designed for an institution-managed AI Foundry: shared GPU compute, persistent project VMs, model-serving endpoints, notebooks, vector databases, experiment tracking, observability, and faculty controls. The strongest education case is predictable shared capacity for whole cohorts, so students can build real systems without personal cloud billing.
What is an AI foundry?Hosted inference, RAG chatbots with optional voice agents, custom LoRA/QLoRA adapters, secure agentic workflows, and GPU FinOps benchmarks require persistent services and measured runtime evidence.
Predictable shared capacity is easier to govern than per-student metered accounts. It supports authentic experimentation while keeping accessibility, access, logs, security, and cost controls visible to faculty and IT.
studio / stem + cs/cis
STEM + CS/CIS collaboration: students build infrastructure for research.
STEM students and faculty bring the scientific question, dataset meaning, and validation criteria. CS/CIS AIINFRA students build the data pipeline, GPU workspace, benchmark or training environment, monitoring, security boundary, cost model, and reproducibility package.
The Well is a public collection of large spatiotemporal physics simulation datasets that can anchor realistic computational-science projects without invented toy data.
Example collaborations
- Physics simulation surrogate modeling on a manageable public dataset subset.
- Research data portal or dashboard for inspecting predictions, errors, and uncertainty.
- Reproducible benchmark comparing local CPU, shared GPU, and managed cloud execution.
- Scientific RAG assistant over papers, lab notes, dataset documentation, and benchmark results.
03 / semester roadmap
Three stackable semesters
Exit after any semester with a coherent, résumé-ready skill set, or continue to the full credential.
Semester 3
Applied Production Systems
216 hrs · 4 courses04 / the ten courses
Every course in the sequence
AIINFRA 100 · Semester 1
Introduction to AI Systems & Infrastructure
The foundation course: AI workload, compute, data & deployment basics
View course AIINFRA 101 · Semester 1Development Environments & Tools
The practical developer toolkit for building AI projects
View course AIINFRA 102 · Semester 1Containerization & GPU-Aware Orchestration
Package, deploy, and scale AI workloads with Docker and Kubernetes
View course AIINFRA 200 · Semester 2Cloud Platforms for AI & CI/CD for ML
Deploy and automate AI workloads on AWS, Azure, and Google Cloud
View course AIINFRA 201 · Semester 2Production Inference Serving & GPU Orchestration
Deploying, Scaling, and Operating LLMs in Production on GPUs
View course AIINFRA 202 · Semester 2Model Adaptation — Fine-Tuning & Quantization
Customize and compress open-weight LLMs on consumer and cloud GPUs
View course AIINFRA 300 · Semester 3Agentic AI & the Model Context Protocol (MCP)
Building, Securing, and Orchestrating Production AI Agents with MCP
View course AIINFRA 301 · Semester 3Production RAG & LLMOps — Observability and Evaluation
Building, Evaluating, and Operating Retrieval-Augmented LLM Systems in Production
View course AIINFRA 302 · Semester 3AI Security, Guardrails & Governance
Securing, Governing, and Cost-Optimizing Production LLM Systems
View course AIINFRA 303 · Semester 3Capstone Project
Design, build, deploy, and present one production-grade AI system
View course05 / careers
What this trains you for
California demand for the roles this certificate builds toward. U.S. Computer Systems Design & related services — employment: 2,374.6K (June 2026, BLS CES).
snapshot · EDD May 2024 employment / Q1 2025 wages
| Data Scientists 15-2051 |
$145,554 | +37.1% | 40,180 |
10th $75,38525th $101,136median $140,51875th $182,30390th $227,088 OEWS current-quarter wage distribution (2025 1st Qtr); the CA median wage column above is the 2024–34 projected median, so the two can legitimately differ. | |||
| Computer & Information Research Scientists 15-1221 |
$163,554 | +26.1% | 7,690 |
10th $85,35825th $108,963median $160,53875th $207,75790th $239,200+ OEWS current-quarter wage distribution (2025 1st Qtr); the CA median wage column above is the 2024–34 projected median, so the two can legitimately differ. | |||
| Software Developers 15-1252 |
$179,292 | +18.4% | 207,260 |
10th $106,77525th $138,484median $175,55575th $218,04990th $239,200+ OEWS current-quarter wage distribution (2025 1st Qtr); the CA median wage column above is the 2024–34 projected median, so the two can legitimately differ. | |||
| Computer & Information Systems Managers 11-3021 |
$221,952 | +16.5% | 82,270 |
10th $129,57925th $168,120median $215,25975th $239,200+90th $239,200+ OEWS current-quarter wage distribution (2025 1st Qtr); the CA median wage column above is the 2024–34 projected median, so the two can legitimately differ. | |||
| Computer Network Architects 15-1241 |
$163,317 | +11.8% | 9,510 |
10th $86,33425th $112,712median $142,11075th $180,19890th $222,189 OEWS current-quarter wage distribution (2025 1st Qtr); the CA median wage column above is the 2024–34 projected median, so the two can legitimately differ. | |||
| Computer Occupations, All Other 15-1299 |
$138,203 | +11.1% | 59,320 |
10th $57,15225th $83,027median $132,24975th $173,73790th $219,241 OEWS current-quarter wage distribution (2025 1st Qtr); the CA median wage column above is the 2024–34 projected median, so the two can legitimately differ. | |||
| Computer Systems Analysts 15-1211 |
$131,295 | +9.7% | 41,500 |
10th $80,19225th $102,523median $131,12975th $165,17090th $202,220 OEWS current-quarter wage distribution (2025 1st Qtr); the CA median wage column above is the 2024–34 projected median, so the two can legitimately differ. | |||
| Network & Computer Systems Administrators 15-1244 |
$109,420 | -4.5% | 11,510 |
10th $69,28325th $86,920median $109,51875th $137,70390th $171,775 OEWS current-quarter wage distribution (2025 1st Qtr); the CA median wage column above is the 2024–34 projected median, so the two can legitimately differ. | |||
Wages refresh live from data.ca.gov when available; otherwise the verified snapshot shown. No occupation-level data comes from the BLS API (BLS provides national context only).
06 / credential & pathway
Credential & pathway
A Stackable pathway — noncredit CDCP Certificate of Completion with credit CTE Certificate of Achievement planning track designed to stack toward credit.
The certificate is organized as three stackable semesters — Foundations, Core Infrastructure, and Applied Production Systems — so a learner can exit after any semester with a coherent, résumé-ready skill set, or continue straight through to the full credential. The noncredit Career Development and College Preparation (CDCP) Certificate of Completion remains the lower-barrier workforce route. The credit CTE Certificate of Achievement remains a planned upper-tier pathway pending local hours-to-units review, COR entry, COCI submission, and curriculum approval.
California labor-market data show the occupations this certificate targets among the fastest-growing in the state: Data Scientists (+37.1%) and Computer & Information Research Scientists (+26.1%) are projected to grow far faster than the statewide all-occupations baseline of 8.8%, and Software Developers — the largest target occupation at 290,800 workers — is projected to add over 200,000 openings through 2034. AI is simultaneously displacing routine technical roles (Network & Computer Systems Administrators is the one target occupation in decline) and creating new demand for workers who can build, deploy, and govern AI systems — precisely the retraining path this certificate provides.