Wirewalk Instituteof AI
bld-201 Intermediate AvailableFlagship

Building with LLM APIs Across Vendors

For developers who want to call OpenAI, Anthropic and Gemini models from Python and TypeScript, and ship code that survives a vendor switch.

Level
Intermediate
Study time
14 hours
Lessons
18 in 6 modules
Exam
20 questions

About this course

Most teams start with one vendor's quickstart and discover the hard parts later: leaked keys, streams that drop, JSON that does not parse, tool loops that never end, a surprise invoice, a 429 storm, and a rewrite when the business wants a second vendor. This course teaches the OpenAI Responses API, the Anthropic Messages API and the Google Gemini Interactions API side by side in Python and TypeScript, with the official SDKs as of October 2026, so you learn the shared concepts and the real differences at the same time.

You will build every capability twice or three times, then learn to put a thin, honest abstraction around them, when OpenAI-compatible endpoints help and when they quietly drop features, and how the same models are reached through Amazon Bedrock, Google Cloud and Microsoft Foundry. The course closes with testing strategies that do not burn tokens and a capstone ticket-triage service with structured output, tools, retries, cost reporting and a one-line vendor switch.

What you will be able to do

  • Configure API keys and cloud credentials for three vendors without committing, logging or shipping secrets to clients
  • Write equivalent text, streaming, structured-output and tool-calling requests against the OpenAI, Anthropic and Gemini SDKs in Python and TypeScript
  • Estimate and measure per-request cost from token usage, and cut it with prompt caching and batch processing
  • Implement error handling that distinguishes retryable from non-retryable failures, with backoff, jitter and Retry-After support
  • Design a thin provider abstraction and decide when OpenAI-compatible endpoints or gateways are appropriate
  • Choose between direct vendor APIs and cloud routes (Bedrock, Google Cloud, Microsoft Foundry) based on identity, data and feature requirements
  • Test LLM integrations with recorded fixtures, contract tests and small live evaluations

Syllabus

6 modules · 18 lessons · about 14 hours

Module 1: Setup, Keys and First Calls

Handle credentials safely and make your first equivalent calls to three vendors in Python and TypeScript.
3 lessons
  1. API Keys, Environments and Secret Hygiene
  2. First Calls in Python: OpenAI, Anthropic and Gemini Side by Side
  3. The Same Calls in TypeScript and Reading the Response Shape

Module 2: Conversations, Parameters and Streaming

Control conversation state, instructions and generation settings, and stream output to users reliably.
3 lessons
  1. Messages, Roles and System Instructions Across Vendors
  2. Generation Parameters, Reasoning Effort and Stop Conditions
  3. Streaming Responses Reliably

Module 3: Structured Outputs, Tools and Multimodal Input

Get schema-valid JSON, run tool-calling loops safely, and send images and PDFs.
3 lessons
  1. Structured Outputs with JSON Schema, Pydantic and Zod
  2. Tool Calling: The Request, Execute, Return Loop
  3. Images and PDFs as Inputs

Module 4: Cost, Caching and Reliability

Measure tokens and cost, use caching and batch discounts, and survive rate limits and outages.
3 lessons
  1. Counting Tokens and Estimating Cost per Request
  2. Prompt Caching and Batch Processing
  3. Errors, Rate Limits and Retries with Backoff

Module 5: Portability and Cloud Routes

Build a thin abstraction, understand compatibility layers, and reach models through AWS, Google Cloud and Azure.
3 lessons
  1. Designing a Thin Provider Abstraction
  2. OpenAI-Compatible Endpoints, Gateways and Their Limits
  3. Cloud Routes: Amazon Bedrock, Google Cloud and Microsoft Foundry

Module 6: Testing and the Capstone Project

Test without burning tokens, then build and harden a multi-vendor ticket triage service.
3 lessons
  1. Testing LLM Integrations Without Burning Tokens
  2. Capstone Part 1: A Multi-Vendor Ticket Triage Service
  3. Capstone Part 2: Hardening, Cost Reporting and a Vendor Switch

Exam and certificate

The exam unlocks once you have completed every lesson. It draws 20 questions from a larger bank, spread across the modules, with the answer options shuffled for each attempt. You have 40 minutes and need 80% to pass. After a failed attempt you can try again after 24 hours, up to 3 attempts in total. Feedback is given per module, never as the answer key.

Passing earns a certificate of completion and assessment from the Wirewalk Institute of AI with a unique ID that anyone can check at institute.wirewalk.com/verify. It is not an accredited qualification or a vendor certification. How certification works.

The Wirewalk Institute of AI is independent. This course teaches publicly documented products and is not affiliated with or endorsed by OpenAI, Anthropic, Google, Amazon Web Services, or Microsoft. Product and company names are trademarks of their respective owners.

More in Building AI Applications

All 14 courses in this track
bld-101 Basic Waitlist

No-Code AI Assistants: Custom GPTs, Projects, Gems and Copilot Studio

A hands-on course on building no-code assistants with the builders already in your workplace tools: ChatGPT projects, custom GPTs and their successor plugins, Claude Projects and Skills, Gemini Gems, and Microsoft 365 Copilot Agent Builder and Copilot Studio. You leave with a tested, documented assistant and a plan to share it safely.

7 h · 17 lessons Join the waitlist
bld-210 Intermediate Available

Retrieval-Augmented Generation Systems

A practitioner's course on retrieval-augmented generation: when to use it, how to ingest, chunk, embed and index documents, how to combine BM25 and vector search with reranking, and how to measure, debug and operate the result. Every module includes working Python sketches.

14 h · 16 lessons See price at checkout
bld-215 Intermediate Waitlist

Embeddings and Vector Databases

A deep dive into the two layers under every semantic search system: embedding models and vector databases. Covers how embeddings are trained and benchmarked, Matryoshka truncation and quantization, ANN index internals (HNSW, IVF-PQ, DiskANN), filtering, multi-tenancy and the operational work of re-embedding and capacity planning.

13 h · 20 lessons Join the waitlist