Skip to content
[ Provider specializations ]

One operating standard.
Four provider lanes.

Learn to make durable architecture decisions first. Then go deep on Anthropic, OpenAI, AWS, or Google Cloud without confusing product knowledge with transferable capability.

Compare the pathwaysInspect the placement method
Specialization routerDesign published
Core

AI
Native

Core → specialize → compare → defend
01[ The classification rule ]

A credential is only as useful as the claim it can honestly carry.

Provider programs use similar language for different kinds of recognition. We preserve those boundaries so a course record never quietly becomes a certification claim.

01

Formal certification

A provider's named certification, governed and issued under that provider's exam rules.

02

Provider credential

A provider-issued credential whose availability and issuance rules are stated by the provider.

03

Completion certificate

Evidence that an eligible course was completed — not automatically a formal certification.

04

Microcredential

A narrower, skills-specific assessment signal that remains distinct from a broad certification.

05

Learning path

Instruction and practice. We make no credential claim unless the official source makes one.

Partner-only

Eligibility depends on a provider partner program. It is never a hidden graduation requirement for a public LockedIn pathway.

Invite-only

Access depends on the provider's selected workspace or rollout. Learners can complete the pathway without it.

02[ The provider lanes ]

Go deep without becoming dependent.

Each lane adds current product practice to the same operating core. Official courses and credentials stay with their issuers; LockedIn evaluates the work produced around them.

Lane 01ANT
[ Claude application architecture ]

Anthropic

An optional implementation lane for production Claude applications, added after the provider-independent operating core.

Best aligned to
  • Solution architects
  • AI engineers
  • Technical leads
LockedIn practice layer
  1. 01Context and tool-interface design
  2. 02Evaluation and failure analysis
  3. 03Production architecture decisions
Official external programs
Completion certificatePublic access

AI Fluency for Builders completion badge

Records completion of Anthropic's builder-focused AI fluency course covering problem framing, acceptance tests, discernment, and human ownership.

How we count it: The public Claude Academy badge can confirm course completion. It is not a proctored Claude Certification and never substitutes for LockedIn mission or field evidence.

Claude Academy · AI Fluency for Builders
Completion certificatePublic access

Claude Platform 101 completion badge

Records completion of Anthropic's foundation course on API requests, agent loops, tools, context management, and managed agents.

How we count it: The public Claude Academy badge can confirm course completion. It is not a proctored Claude Certification and never substitutes for LockedIn mission or field evidence.

Claude Academy · Claude Platform 101
Completion certificatePublic access

Building with the Claude API completion badge

Records completion of Anthropic's extended Claude API course across prompting, tools, retrieval, agents, MCP, evaluation, and production patterns.

How we count it: The public Claude Academy badge can confirm course completion. It is not a proctored Claude Certification and never substitutes for LockedIn mission or field evidence.

Claude Academy · Building with the Claude API
Completion certificatePublic access

Claude Code 101 completion badge

Records completion of Anthropic's introductory Claude Code course for core terminal and engineering workflows.

How we count it: The public Claude Academy badge can confirm course completion. It is not a proctored Claude Certification and never substitutes for LockedIn mission or field evidence.

Claude Academy · Claude Code 101
Completion certificatePublic access

Claude Code in Action completion badge

Records completion of Anthropic's longer-running Claude Code workflow course covering steering, configuration, automation, and verification.

How we count it: The public Claude Academy badge can confirm course completion. It is not a proctored Claude Certification and never substitutes for LockedIn mission or field evidence.

Claude Academy · Claude Code in Action
Completion certificatePublic access

Introduction to Model Context Protocol completion badge

Records completion of Anthropic's MCP foundation course for servers, clients, tools, resources, and prompts.

How we count it: The public Claude Academy badge can confirm course completion. It is not a proctored Claude Certification and never substitutes for LockedIn mission or field evidence.

Claude Academy · Introduction to Model Context Protocol
Completion certificatePublic access

Model Context Protocol: Advanced Topics completion badge

Records completion of Anthropic's advanced MCP course covering sampling, notifications, roots, transports, and deployment tradeoffs.

How we count it: The public Claude Academy badge can confirm course completion. It is not a proctored Claude Certification and never substitutes for LockedIn mission or field evidence.

Claude Academy · Model Context Protocol: Advanced Topics
Completion certificatePublic access

Claude with Amazon Bedrock completion badge

Records completion of Anthropic's AWS deployment course for Claude applications, tools, retrieval, agents, and production operation on Bedrock.

How we count it: The public Claude Academy badge can confirm course completion. It is not a proctored Claude Certification and never substitutes for LockedIn mission or field evidence.

Claude Academy · Claude with Amazon Bedrock
Completion certificatePublic access

Claude with Google Cloud's Vertex AI completion badge

Records completion of Anthropic's Google Cloud deployment course for Claude applications, tools, retrieval, agents, and production operation on Vertex AI.

How we count it: The public Claude Academy badge can confirm course completion. It is not a proctored Claude Certification and never substitutes for LockedIn mission or field evidence.

Claude Academy · Claude with Google Cloud's Vertex AI
Formal certificationPartner-only

Claude Certified Associate: Foundations

Validates practical, everyday use of Claude for people working on Claude-related projects, including consultants, project leads, and business or technical practitioners.

How we count it: Count only when Anthropic issues the credential after its proctored, identity-verified exam; exams are available to Claude Partner Network members and this credential is not a LockedIn pathway prerequisite.

Claude · role-based certification announcement
Formal certificationPartner-only

Claude Certified Developer: Foundations

For engineers building applications with Claude, with coverage that includes the Claude API, tool use, and agent development.

How we count it: Count only when Anthropic issues the credential after its proctored, identity-verified exam; exams are available to Claude Partner Network members and this credential is not a LockedIn pathway prerequisite.

Claude · role-based certification announcement
Formal certificationPartner-only

Claude Certified Architect: Foundations

For solution architects who design and build agent systems with Claude.

How we count it: Count only when Anthropic issues the credential after its proctored, identity-verified exam; exams are available to Claude Partner Network members and this credential is not a LockedIn pathway prerequisite.

Claude · role-based certification announcement
Formal certificationPartner-only

Claude Certified Architect: Professional

Anthropic's advanced credential for large-enterprise integration architecture, governance, and evaluation.

How we count it: Count only when Anthropic issues the credential after its proctored, identity-verified exam; Anthropic says professional paths advance from a foundation-level certification, and exams are available to Claude Partner Network members.

Claude · role-based certification announcement
Lane 02OAI
[ AI applications and agent workflows ]

OpenAI

An optional build lane for OpenAI application patterns, agent workflows, evaluation, and production operation.

Best aligned to
  • Application developers
  • Agent engineers
  • AI product teams
LockedIn practice layer
  1. 01Tool-using application workflows
  2. 02Agent review and evaluation
  3. 03Production reliability decisions
Official external programs
Completion certificatePublic access

OpenAI Academy courses

Free, self-paced courses in AI foundations, applied workflows, and agents are available to people with a ChatGPT account.

How we count it: OpenAI states that Academy completion certificates confirm course completion and are not OpenAI Certifications or formal OpenAI credentials.

OpenAI Help Center · Academy courses
Provider credentialInvite-only

OpenAI Certified app

A Coursera-powered learning experience in ChatGPT with assessments and eligible OpenAI-issued credentials distributed through Credly.

How we count it: OpenAI currently limits access to invited ChatGPT Enterprise and Edu workspaces; this restricted credential is never required to complete the LockedIn pathway.

OpenAI Help Center · Certified app
Learning pathPublic access

AI app development: concept to production

An official developer track organized around taking an AI application from concept through production.

How we count it: Used as current technical reference material; the public track page does not by itself establish a certification claim.

OpenAI Developers · learning track
Lane 03AWS
[ Production AI on AWS ]

Amazon Web Services

An optional cloud implementation lane spanning AI fluency, production ML, generative AI applications, and agentic operations on AWS.

Best aligned to
  • Cloud engineers
  • ML engineers
  • Generative AI developers
LockedIn practice layer
  1. 01Bedrock-centered application architecture
  2. 02ML deployment and operations
  3. 03Security, governance, and production troubleshooting
Official external programs
Formal certificationPublic access

AWS Certified AI Practitioner

A foundational certification for AI, machine learning, generative AI concepts, and AWS use cases.

How we count it: Useful as an independently issued knowledge signal; it does not replace evidence that a learner can build or operate a system.

AWS Certification · AI Practitioner
Formal certificationPublic access

AWS Certified Machine Learning Engineer — Associate

A role-based certification focused on implementing ML workloads in production and operationalizing them.

How we count it: Adds provider-specific ML operations evidence alongside the pathway's original artifacts and defense.

AWS Certification · ML Engineer — Associate
Formal certificationPublic access

AWS Certified Generative AI Developer — Professional

A professional certification for integrating foundation models into applications and business workflows in production environments using AWS technologies.

How we count it: A broad provider certification; LockedIn separately evaluates the implementation artifact and its operating evidence.

AWS Certification · AIP-C01 exam guide
MicrocredentialPublic access

AWS Agentic AI Demonstrated

AWS describes this microcredential as a practical implementation assessment in a provisioned AWS environment.

How we count it: A narrower hands-on signal that complements, but is distinct from, an AWS Certification.

AWS Training and Certification · portfolio announcement
Lane 04GCP
[ Generative AI and ML on Google Cloud ]

Google Cloud

An optional cloud lane for generative AI leadership, application development, model lifecycle work, and production ML on Google Cloud.

Best aligned to
  • AI leaders
  • Application developers
  • ML engineers
LockedIn practice layer
  1. 01Gemini and Google Cloud solution decisions
  2. 02Model deployment and MLOps
  3. 03Responsible AI and lifecycle controls
Official external programs
Formal certificationPublic access

Google Cloud Generative AI Leader

A business-level certification covering generative AI fundamentals, Google Cloud offerings, output improvement, and adoption strategy.

How we count it: Adds an independently issued leadership signal; it is not treated as evidence of hands-on engineering by itself.

Google Cloud Certification · Generative AI Leader
Formal certificationPublic access

Google Cloud Professional Machine Learning Engineer

A professional certification for building, evaluating, productionizing, and optimizing AI solutions with Google Cloud.

How we count it: Adds a broad provider certification; original deployment evidence remains the decisive pathway signal.

Google Cloud Certification · Professional ML Engineer
Learning pathPublic access

Advanced: Generative AI for Developers

A Google Cloud-managed technical learning path for application developers, machine learning engineers, and data scientists.

How we count it: Treated as supplemental instruction; a learning-path record is not presented as a Google Cloud certification.

Google Skills · learning path
Learning pathPublic access

Deploy and Manage Generative AI Models

A Google Cloud-managed path focused on the generative AI model lifecycle, deployment, monitoring, and a hands-on Vertex AI lab.

How we count it: Used to deepen provider-specific MLOps practice without being restated as a certification.

Google Skills · learning path
03[ Evidence weighting ]

The provider badge is additive. The work remains decisive.

We do not collapse unlike evidence into a cosmetic points score. The record states what was built, what was compared, who verified it, and what an external provider independently issued.

01Decisive

Applied capability

Original artifacts, evaluation results, operating logs, and a live defense determine whether the learner can do the work.

Verified by · LockedIn Labs
02Required

Provider-comparison capstone

Once this roadmap capstone is authored and activated, the learner must prove judgment across two provider stacks, including the tradeoffs, controls, and migration boundary between them.

Verified by · LockedIn Labs
03Additive

External credential

An active provider-issued certification, credential, or microcredential strengthens the record but never substitutes for proof-of-work.

Verified by · The named provider
04[ Roadmap specification · provider-comparison capstone ]

One brief. Two stacks. A defensible choice.

Build or reproduce the same bounded enterprise workflow across two approved provider stacks, then defend which architecture should ship and where portability ends. This is the published design contract; it is not represented as an available lab until its modules, rubrics, permission boundaries, and delivery schedule are activated.

Acceptance contract
  1. 01Use the same acceptance criteria and held-out evaluation set for both implementations
  2. 02Map model, tool, data, identity, and human-authorization boundaries
  3. 03Compare quality, latency, operating cost, observability, safety, and failure recovery
  4. 04Document a migration path and the dependencies that cannot be made portable
Evidence package
OUT · 01

Two runnable implementation slices

OUT · 02

Evaluation harness and result record

OUT · 03

Provider tradeoff and portability matrix

OUT · 04

Architecture decision record and live defense

Provider knowledge, without provider lock-in

Bring your target role. We will map the core and the right depth lane.

Design an enterprise pathway
[ Independent program notice ]

LockedIn Labs is independent and is not endorsed, sponsored, authorized, or certified by Anthropic, OpenAI, Amazon Web Services, or Google Cloud. Provider names and marks belong to their respective owners. External programs are governed and issued by the named provider, and access or availability may change. We link to official material for supplemental study; we do not reproduce provider courseware, proprietary assessment content, or exam questions. Curriculum design is published here; delivery availability appears only when listed on the cohort schedule.