The Exam Room

29 Apr 2026 · 16 min read

Exam Room · GenAI

Picking the AWS AI Service Tier for Each Feature

A product manager with no ML background has been told to add AI to a SaaS product, and has heard of Bedrock, SageMaker, Comprehend, Translate, Textract, Rekognition. AWS has three different shapes of AI offering, and the shortest path depends entirely on whether a ready-made service already does the job.

Read article 04 May 2026 · 19 min read

Exam Room · GenAI

How to Take a Foundation Model from Pick to Production Endpoint

A product team wants a chatbot that summarises support tickets. They have the tickets, a cloud account, and no ML background. Somebody says 'use a foundation model'. Between that sentence and a working endpoint sit roughly seven distinct stages, each with its own AWS service and its own decisions. Picking the model is the easy part; the real work is figuring out which stages this team can skip, which they absolutely cannot, and what AWS gives them at each step.

Read article 06 May 2026 · 19 min read

Exam Room · GenAI

Choosing Between Prompting, RAG, and Fine-Tuning

A legal-ops team wants a model that answers questions about their 4,000 in-house contract templates. The first prototype, a plain Claude call with the question in the prompt, hallucinates clause numbers. Someone suggests fine-tuning; someone else suggests RAG. They solve different problems, so 'which is better' is the wrong frame; what matters is which problem the team actually has, and what each adaptation technique costs in time, data, and recurring spend.

Read article 11 May 2026 · 20 min read

Exam Room · GenAI

Grounding a Chatbot in Your Own PDFs

A facilities team has 600 PDFs, equipment manuals, safety procedures, maintenance schedules, sitting on a SharePoint drive. Engineers want a chatbot that answers 'how do I reset the chiller on floor 4?' in seconds instead of a ten-minute PDF hunt. Retrieval-augmented generation can do this; whether it does it well depends on what the corpus actually looks like, what kinds of questions the engineers really ask, and which configuration knobs decide whether the answers are any good once a managed service is on the table.

Read article 13 May 2026 · 18 min read

Exam Room · GenAI

Forecasting Without Writing Python

A category manager has 18 months of weekly sales data for 400 SKUs and a deadline to forecast next quarter. She doesn't code. The ML team is booked until Q3. The ask is a tool that lets her build a forecast herself, importable, reviewable, explainable, without waiting for engineering. Which AWS box she clicks matters less than what kind of problem this actually is, what features of the data can honestly feed into a model, and what the business user has to understand for the output to be defensible when finance asks ''why this number?''.

Read article 18 May 2026 · 21 min read

Exam Room · GenAI

How to Make a Bedrock Chatbot Audit-Ready with Guardrails and Watermarks

A fintech ships a customer-facing chatbot on Bedrock. Legal asks: can it give financial advice? Risk asks: can it leak customer account numbers? Compliance asks: if an auditor requests proof a response came from our model, can we demonstrate it? Three questions, three different controls, all of them Bedrock-native. The controls exist; the work is matching the right one to each question and figuring out what the shape of a 'responsible AI' configuration actually looks like when the auditor arrives.

Read article 20 May 2026 · 19 min read

Exam Room · GenAI

Choosing Between Chains, Retrieval, and Agents for a GenAI Assistant

A product manager wants a 'GenAI assistant' for internal operations, something that can answer questions, look up customer records, draft emails, and file Jira tickets. Three architectural patterns keep coming up: chains, retrieval, and agents. They sound similar, they all use foundation models, and teams routinely reach for the most elaborate one when a simpler pattern would do. There's no single 'best' here; what matters is which one fits each piece of the assistant's workload, and when elaboration costs more than it earns.

Read article 20 May 2026 · 14 min read

Exam Room · GenAI

How to Match Inference Parameters to the Job at Hand

Three features share one foundation model. An email-to-JSON extractor that sometimes wraps its JSON in friendly prose. A support-reply drafter whose 'regenerate' button produces the same draft five times. A tagline brainstormer that returns ten near-identical suggestions. Nobody has touched the inference parameters; everything runs on the defaults. Temperature, top-p, top-k, maximum tokens, and stop sequences are the knobs that decide what kind of text comes back, and each of these three jobs wants them set differently.

Read article 25 May 2026 · 21 min read

Exam Room · GenAI

Choosing Between SageMaker, Bedrock, and Purpose-Built AI APIs

A platform team has five AI-shaped requests landing in a single sprint: transcribe call centre audio, detect anomalies in sensor data, extract text from scanned forms, summarise customer emails, and detect faces in CCTV. Someone has already typed 'use SageMaker' into three design docs. Someone else insists Bedrock is the answer. A third voice mutters about purpose-built services. AWS has at least three answers to every AI problem, so there's no single platform that wins; what matters is how to tell which layer of the stack each request lands on, and what that choice costs in time, money, and flexibility.

Read article 25 May 2026 · 10 min read

Exam Room · GenAI

Choosing Responsible-AI Controls for a Loan-Approval Model

A regional lender puts a machine-learning model in charge of personal-loan decisions, and three uncomfortable questions arrive within a month. A declined applicant asks 'why?'. The risk team asks whether the model treats different groups of applicants fairly. The regulator asks for documentation and proof that a human reviews the borderline cases. None of these is answered by accuracy metrics, and none is answered by a single product. AWS ships a toolkit for exactly this, and the work is matching each question to the control that actually answers it.

Read article 27 May 2026 · 15 min read

Exam Room · Advanced GenAI

Picking a Bedrock Model for High-Volume RAG

A million LLM requests a day, peaking at thirty per second, split across US and EU customers, with a P99 first-token target under 1.5 seconds and real reasoning over retrieved context. Bedrock has seven model families and four ways to buy capacity. Most of the landscape falls away once you name what actually decides it, and the real trick is what you do *after* you've picked the model.

Read article 01 Jun 2026 · 17 min read

Exam Room · Advanced GenAI

How to Build a Citations-Required RAG Over 50K Internal Documents

Fifty thousand internal documents, five gigabytes of text, weekly churn, a three-second latency budget, per-user access control, and a citation in every single answer. The RAG landscape on Bedrock is bigger than one product and the interesting part of the design is what falls away once you name the five things that actually decide it.

Read article 03 Jun 2026 · 18 min read

Exam Room · Advanced GenAI

Combining RAG and Fine-Tuning for a Legal Contract Assistant

A legal-tech team wants a contract review assistant that understands two hundred thousand past matters, speaks in the firm's voice with clause-by-section citations, and refuses anything off-domain. A hundred thousand dollars, three months. RAG, fine-tuning, and continued pre-training each solve a different half of that sentence; the interesting answer is which two to pick, not which one.

Read article 08 Jun 2026 · 16 min read

Exam Room · Advanced GenAI

Designing Short-Term and Long-Term Memory for a Bedrock Chat Assistant

A customer-support assistant where the average conversation runs fifteen turns before it resolves, and returning users pick up two weeks later expecting the bot to remember they've been waiting on a refund. Two memory problems in one product (what's live in the current conversation and what persists across visits) and four plausible ways to build it. Bedrock Agents' built-in memory handles one half cleanly; the other half is where teams reach for DynamoDB or a knowledge base and get it wrong.

Read article 15 Jun 2026 · 17 min read

Exam Room · Advanced GenAI

Configuring Bedrock Guardrails for PII, Topics, and Grounding

A consumer-facing chatbot on Bedrock has passed every red-team round on the obvious harms (no weapons, no hate, no CSAM) and is still shipping embarrassments: a card number pasted by one user echoing back in a reply, the bot cheerfully comparing the company's product with a named competitor, and a hallucinated policy line that nobody in the building wrote. Five different filter jobs wrap the same Bedrock invocation, and Guardrails is the one surface that does all five without five Lambdas.

Read article 17 Jun 2026 · 15 min read

Exam Room · Advanced GenAI

Spreading Bedrock Load with Cross-Region Inference Profiles

A Bedrock-backed SaaS serving US, EU and APAC customers is hitting regional quota in us-east-1 during peak while the same model sits idle in eu-west-1. The team wants to spread load without fracturing the product into three regional deployments. Three letters on the front of the model ID do the job, provided the model supports it and the geography fits the customer.

Read article 19 Jun 2026 · 16 min read

Exam Room · Advanced GenAI

Building RAG When the Source Documents Change Daily

A support assistant that has to answer from a product manual which the product team edits weekly, a pricing sheet that changes at month-end, and an operational runbook that mutates hourly. The base model doesn't know any of it, and fine-tuning won't keep up. Retrieval is the answer; the question is how much of the retrieval plumbing we want to own, and Bedrock Knowledge Bases, a LangChain stack, and a hand-rolled pipeline each put the lines in different places.

Read article 22 Jun 2026 · 15 min read

Exam Room · Advanced GenAI

How to Wire an LLM to Side-Effecting Actions with Bedrock Agents

An assistant that has to look up a customer's subscription, pause it, refund a charge, and email confirmation. Not just answer, act. The glue between a language model and the rest of our systems is a solved problem three different ways: Bedrock Agents, a LangChain agent loop, or a hand-written tool router. Each of them handles tool definition, invocation, and error recovery, but they put the guardrails in very different places.

Read article 24 Jun 2026 · 14 min read

Exam Room · Advanced GenAI

Picking a Vector Store for Bedrock RAG

Twelve million embedding vectors, a 50ms retrieval budget, hybrid queries that mix keyword and semantic, and a bill that should not double the Bedrock spend on its own. OpenSearch Serverless, Aurora with pgvector, and Pinecone Serverless all serve the same shape of query, but their pricing curves, operational shapes, and query surfaces diverge the moment the corpus grows beyond demo scale.

Read article 29 Jun 2026 · 17 min read

Exam Room · Advanced GenAI

How to Manage Prompts Across Thirty Services on Bedrock

One prompt scattered across thirty services, no versioning, no tests, drift between the copy in the code and the copy in the docs, a silent regression when somebody changed 'concise' to 'brief' and retention on one response tanked. Prompt engineering at a hundred callers isn't prose discipline, it's configuration management. Bedrock Prompt Management, Git-backed templates, and parameterised prompts each solve a slice of the same problem.

Read article 01 Jul 2026 · 13 min read

Exam Room · Advanced GenAI

Evaluating LLM Output With Bedrock Eval Jobs

Two thousand historical support tickets, a summarisation prompt, a new model candidate, and a product manager asking whether switching would hurt quality. Bedrock evaluation jobs offer automated scoring, human review through Ground Truth workflows, and model comparison side by side, but they answer different questions, and getting the correct number out of the correct job matters more than running more of them.

Read article 03 Jul 2026 · 15 min read

Exam Room · Advanced GenAI

How to Build a Multi-Modal Bedrock Assistant for Insurance Claims

A claims-processing assistant that reads a scanned invoice, listens to a voicemail, answers the customer's question in plain text, and, if asked, reads it back. Four modalities, one conversation. The model choice, the orchestration shape, and the ways different inputs fail each push the architecture in different directions, and the naive 'just use a multi-modal model' misses half of where the real work is.

Read article 06 Jul 2026 · 15 min read

Exam Room · Advanced GenAI

How to Cut a Bedrock Bill Without Hurting Quality

A Bedrock bill that doubled in two months, a product roadmap that blames the retrieval service, a finance partner who'd like a straight answer. The cheapest token is the one you don't send; the next cheapest is the one you send to the correct model. Model routing, prompt compression, cached retrieval, and provisioned throughput each solve a different slice of the cost problem, and none of them is the silver bullet.

Read article 08 Jul 2026 · 15 min read

Exam Room · Advanced GenAI

Importing Custom Weights into Bedrock

A research team has fine-tuned an open-weights model for medical-notes summarisation on a private SageMaker cluster. The resulting weights live in S3; the production runtime wants Bedrock's ergonomics. Custom model import bridges that gap, but it only works for certain base architectures, comes with throughput minimums, and quietly changes the cost model compared to on-demand foundation models.

Read article