<?xml version="1.0" encoding="UTF-8"?>
<rss version="2.0"><channel><title>FDE Times</title><link>https://fdetimes.net/en/</link><description>News, analysis, guides, a learning roadmap and jobs for Forward Deployed Engineers.</description><language>en</language><item><title>Build a webhook receiver that withstands forged signatures, replays and duplicate events</title><link>https://fdetimes.net/en/guides/secure-webhook-receiver-signatures-replay-duplicates/</link><guid>https://fdetimes.net/en/guides/secure-webhook-receiver-signatures-replay-duplicates/</guid><pubDate>Fri, 09 Oct 2026 07:38:05 GMT</pubDate><description>Stripe can resend the same event for up to three days, with a new signature each time. A time window cannot stop those duplicates, so your receiver needs a separate line of defence.</description></item><item><title>Reconciliation: only say &quot;it works&quot; once your numbers match the client&apos;s books</title><link>https://fdetimes.net/en/guides/data-reconciliation-before-saying-it-works/</link><guid>https://fdetimes.net/en/guides/data-reconciliation-before-saying-it-works/</guid><pubDate>Fri, 09 Oct 2026 07:38:01 GMT</pubDate><description>An all-green dashboard does not mean the numbers are right. The client&apos;s chief accountant decides whether they are right, and she will check them against her own ledger.</description></item><item><title>DPAs, data terms and procurement: what FDEs need to know so projects don&apos;t stall in legal</title><link>https://fdetimes.net/en/guides/dpa-data-terms-procurement-for-fdes/</link><guid>https://fdetimes.net/en/guides/dpa-data-terms-procurement-for-fdes/</guid><pubDate>Fri, 09 Oct 2026 07:37:49 GMT</pubDate><description>A pipeline that works can still stop because legal emails one question: where does the customer&apos;s data go, who holds it, and for how long? That only becomes a problem if you don&apos;t already have the answer.</description></item><item><title>Read Marty Cagan&apos;s Inspired and stop running a feature factory for your customers</title><link>https://fdetimes.net/en/books-courses/inspired-marty-cagan-fde-feature-factory/</link><guid>https://fdetimes.net/en/books-courses/inspired-marty-cagan-fde-feature-factory/</guid><pubDate>Fri, 09 Oct 2026 07:36:26 GMT</pubDate><description>Marty Cagan writes for product managers, but his thinking names the trap FDEs fall into most often: shipping every feature on the customer&apos;s list without knowing what problem has been solved.</description></item><item><title>Hugging Face&apos;s MCP Course: FDEs should study Unit 3 closely and move quickly through Unit 2</title><link>https://fdetimes.net/en/books-courses/hugging-face-mcp-course-review/</link><guid>https://fdetimes.net/en/books-courses/hugging-face-mcp-course-review/</guid><pubDate>Fri, 09 Oct 2026 07:33:42 GMT</pubDate><description>The free course has three parts. Only one of them resembles what an FDE actually does on a customer site.</description></item><item><title>How many clients can one FDE carry? The answer depends on phase and tooling, not on a single number</title><link>https://fdetimes.net/en/analysis/how-many-clients-can-an-fde-handle/</link><guid>https://fdetimes.net/en/analysis/how-many-clients-can-an-fde-handle/</guid><pubDate>Fri, 09 Oct 2026 07:31:58 GMT</pubDate><description>No company has published this figure. The playbooks still agree on one point: how many accounts you can carry depends on the phase each account is in and on how much of the work has been turned into tools.</description></item><item><title>Metabase: self-service dashboards on a client&apos;s database, built in an afternoon</title><link>https://fdetimes.net/en/tools/metabase-self-service-dashboards-client-database/</link><guid>https://fdetimes.net/en/tools/metabase-self-service-dashboards-client-database/</guid><pubDate>Fri, 09 Oct 2026 07:31:06 GMT</pubDate><description>Installing Metabase takes minutes. The hard part is the three decisions that follow: the database account, group permissions and the embedding method. Get any one of them wrong and a polished afternoon demo turns into next week&apos;s incident.</description></item><item><title>Writing your first eval with Inspect AI: dataset, solver, scorer and reading the log</title><link>https://fdetimes.net/en/guides/first-eval-inspect-ai-guide/</link><guid>https://fdetimes.net/en/guides/first-eval-inspect-ai-guide/</guid><pubDate>Fri, 09 Oct 2026 06:38:17 GMT</pubDate><description>With just three questions about a returns policy, a string-matching scorer can grade a model 3/3 when it actually got only 1/3 right. Only the log file will show you that.</description></item><item><title>Sub-agent, skill or MCP server: where a client&apos;s workflow belongs</title><link>https://fdetimes.net/en/guides/sub-agent-skill-or-mcp-server/</link><guid>https://fdetimes.net/en/guides/sub-agent-skill-or-mcp-server/</guid><pubDate>Fri, 09 Oct 2026 06:37:13 GMT</pubDate><description>An agent can fail at a client site even when the model is perfectly capable. Often the cause is simply that business process knowledge was put in the wrong place: the system prompt, a tool description, or a sub-agent nobody needed.</description></item><item><title>Saying “no” to a bad design without losing the client</title><link>https://fdetimes.net/en/guides/how-to-say-no-to-client-design/</link><guid>https://fdetimes.net/en/guides/how-to-say-no-to-client-design/</guid><pubDate>Fri, 09 Oct 2026 06:36:34 GMT</pubDate><description>A refusal that comes with a reason and an alternative, and gets written down, usually keeps a client’s trust better than a yes.</description></item><item><title>AWS Cloud Practitioner Essentials: 13 modules that teach you to talk to a client&apos;s IT team</title><link>https://fdetimes.net/en/books-courses/aws-cloud-practitioner-essentials-for-fdes/</link><guid>https://fdetimes.net/en/books-courses/aws-cloud-practitioner-essentials-for-fdes/</guid><pubDate>Fri, 09 Oct 2026 06:34:57 GMT</pubDate><description>AWS&apos;s beginner cloud course will not turn you into an infrastructure architect. It will help you keep up in your first meeting on a client&apos;s systems.</description></item><item><title>Does needing FDEs mean the product isn&apos;t finished? Follow the custom code</title><link>https://fdetimes.net/en/analysis/fde-product-unfinished-custom-code/</link><guid>https://fdetimes.net/en/analysis/fde-product-unfinished-custom-code/</guid><pubDate>Fri, 09 Oct 2026 06:34:14 GMT</pubDate><description>Investors disagree about whether forward deployed engineering is a deliberate strategy or a sign that a company is turning into a consultancy. The answer lies in what happens to the code after each deployment.</description></item><item><title>Sentry for FDEs: fixing bugs on a client site you are not at</title><link>https://fdetimes.net/en/tools/sentry-for-forward-deployed-engineers/</link><guid>https://fdetimes.net/en/tools/sentry-for-forward-deployed-engineers/</guid><pubDate>Fri, 09 Oct 2026 06:30:19 GMT</pubDate><description>You have no SSH access, no logs, and a client who just writes &quot;it&apos;s broken&quot;. Sentry tells you what failed, where, and in which release.</description></item><item><title>How to agree with the client on what ‘success’ means before you write any AI code</title><link>https://fdetimes.net/en/guides/define-ai-acceptance-criteria-with-clients/</link><guid>https://fdetimes.net/en/guides/define-ai-acceptance-criteria-with-clients/</guid><pubDate>Fri, 09 Oct 2026 05:36:09 GMT</pubDate><description>The client wants a system that is ‘always right’. Any model will sometimes be wrong. If you want the project signed off, agree on how many errors the client will accept in the very first week.</description></item><item><title>Berkeley RDI&apos;s Agentic AI course: what to study before you take agents into the enterprise</title><link>https://fdetimes.net/en/books-courses/berkeley-agentic-ai-mooc-for-fdes/</link><guid>https://fdetimes.net/en/books-courses/berkeley-agentic-ai-mooc-for-fdes/</guid><pubDate>Fri, 09 Oct 2026 05:34:10 GMT</pubDate><description>For a Forward Deployed Engineer, the first lecture to watch in this course is about running agents in production, not about models.</description></item><item><title>When the word &quot;resolved&quot; decides an agent&apos;s revenue</title><link>https://fdetimes.net/en/analysis/outcome-based-pricing-ai-agents-fde/</link><guid>https://fdetimes.net/en/analysis/outcome-based-pricing-ai-agents-fde/</guid><pubDate>Fri, 09 Oct 2026 05:33:56 GMT</pubDate><description>Sierra and Intercom charge for outcomes, so defining what counts as &quot;resolved&quot; is a contract term and a measurement problem. It is also everyday work for engineers who work alongside customers.</description></item><item><title>Working Effectively with Legacy Code: the 2004 book that helps FDEs add AI features to untested code</title><link>https://fdetimes.net/en/books-courses/working-effectively-with-legacy-code-fde/</link><guid>https://fdetimes.net/en/books-courses/working-effectively-with-legacy-code-fde/</guid><pubDate>Fri, 09 Oct 2026 05:30:56 GMT</pubDate><description>The customer system you are about to wire an LLM into almost certainly has no tests. Michael Feathers explained how to handle that more than twenty years ago.</description></item><item><title>LiteLLM Proxy: one shared LLM gateway, a virtual key per client, separate budgets and bills</title><link>https://fdetimes.net/en/tools/litellm-proxy-virtual-keys-budgets/</link><guid>https://fdetimes.net/en/tools/litellm-proxy-virtual-keys-budgets/</guid><pubDate>Fri, 09 Oct 2026 05:29:04 GMT</pubDate><description>When three clients share one OpenAI account, someone will ask who spent what sooner than you expect. A self-hosted gateway is the cleanest answer.</description></item><item><title>From thumbs-down to eval set: turning user complaints into test cases</title><link>https://fdetimes.net/en/guides/thumbs-down-feedback-to-eval-test-cases/</link><guid>https://fdetimes.net/en/guides/thumbs-down-feedback-to-eval-test-cases/</guid><pubDate>Fri, 09 Oct 2026 04:38:00 GMT</pubDate><description>When a client sends you a file full of &quot;dissatisfied&quot; clicks, the first job is to read the traces and label them, not to count the clicks.</description></item><item><title>When the shift supervisor doesn&apos;t trust AI: how an FDE overcomes resistance in a pilot</title><link>https://fdetimes.net/en/guides/frontline-resistance-ai-pilot-fde/</link><guid>https://fdetimes.net/en/guides/frontline-resistance-ai-pilot-fde/</guid><pubDate>Fri, 09 Oct 2026 04:37:16 GMT</pubDate><description>A pilot starts to fail when the model works but users quietly ignore it. The fix is on the shop floor, next to the people using the tool. It is not in the code.</description></item><item><title>Messy client Excel files: count the missing cells before pandas adds them up</title><link>https://fdetimes.net/en/guides/clean-messy-client-excel-pandas/</link><guid>https://fdetimes.net/en/guides/clean-messy-client-excel-pandas/</guid><pubDate>Fri, 09 Oct 2026 04:34:44 GMT</pubDate><description>A single &quot;not yet&quot; in a money column can make reported revenue lower than it really is, and nobody notices, because pandas treats a missing cell as zero.</description></item><item><title>The FDE CV: drop the framework list, state deployment results with a baseline</title><link>https://fdetimes.net/en/analysis/fde-cv-measured-results-not-frameworks/</link><guid>https://fdetimes.net/en/analysis/fde-cv-measured-results-not-frameworks/</guid><pubDate>Fri, 09 Oct 2026 04:33:54 GMT</pubDate><description>People hiring forward deployed engineers read a CV looking for a result they can verify, so a line reading “LangChain, Kafka, Kubernetes” tells them almost nothing.</description></item><item><title>Michael Nygard&apos;s Release It!: the bedside book for FDEs who live in a client&apos;s production</title><link>https://fdetimes.net/en/books-courses/release-it-nygard-production-stability-fde/</link><guid>https://fdetimes.net/en/books-courses/release-it-nygard-production-stability-fde/</guid><pubDate>Fri, 09 Oct 2026 04:32:20 GMT</pubDate><description>A client API that suddenly hangs for 30 seconds per call can drain your service&apos;s threads in just 5 seconds. This 376-page book shows how to break that chain of failures at design time.</description></item><item><title>Hugging Face&apos;s LLM Course: how the first three chapters take an FDE from pipeline() to fine-tuning</title><link>https://fdetimes.net/en/books-courses/hugging-face-llm-course-first-three-chapters/</link><guid>https://fdetimes.net/en/books-courses/hugging-face-llm-course-first-three-chapters/</guid><pubDate>Fri, 09 Oct 2026 04:31:09 GMT</pubDate><description>The course is free and available in many languages. Its first three chapters can be read as the order of work at a client site: set a baseline first, inspect the data next, fine-tune last.</description></item><item><title>Retool for FDEs: build internal tools in hours, and know when to go back to React</title><link>https://fdetimes.net/en/tools/retool-internal-tools-when-to-use-react/</link><guid>https://fdetimes.net/en/tools/retool-internal-tools-when-to-use-react/</guid><pubDate>Fri, 09 Oct 2026 04:30:55 GMT</pubDate><description>Retool can get you a demo in your first week on a client site, but only if you know in advance where it stops paying off.</description></item><item><title>Building an FDE portfolio repo: a simple agent, evals built from real failures, a README and a 3-minute video</title><link>https://fdetimes.net/en/guides/fde-portfolio-repo-agent-evals-readme/</link><guid>https://fdetimes.net/en/guides/fde-portfolio-repo-agent-evals-readme/</guid><pubDate>Fri, 09 Oct 2026 03:03:30 GMT</pubDate><description>A polished demo only proves the agent got it right once. A repo worth sending shows where the agent fails and how often.</description></item><item><title>When AI writes code at scale, FDEs are paid to own the outcome</title><link>https://fdetimes.net/en/analysis/ai-writes-code-fde-owns-outcome/</link><guid>https://fdetimes.net/en/analysis/ai-writes-code-fde-owns-outcome/</guid><pubDate>Fri, 09 Oct 2026 02:59:21 GMT</pubDate><description>Salesforce has written into its FDE job listing that code may come from AI tools, but the person accountable for the result must still be an engineer.</description></item><item><title>The Pyramid Principle: the book that teaches FDEs to write so client executives get it from the first sentence</title><link>https://fdetimes.net/en/books-courses/pyramid-principle-minto-writing-for-fdes/</link><guid>https://fdetimes.net/en/books-courses/pyramid-principle-minto-writing-for-fdes/</guid><pubDate>Fri, 09 Oct 2026 02:58:24 GMT</pubDate><description>At consulting firms, a draft can come back with a single instruction: &quot;make it Minto&quot;. Forward deployed engineers, who write for client leadership, would do well to pick up the same habit.</description></item><item><title>Pydantic AI: typed tools and evals let FDEs switch models without gambling</title><link>https://fdetimes.net/en/tools/pydantic-ai-typed-tools-evals/</link><guid>https://fdetimes.net/en/tools/pydantic-ai-typed-tools-evals/</guid><pubDate>Fri, 09 Oct 2026 02:55:04 GMT</pubDate><description>In Pydantic AI, switching models takes a one-string edit. To show a client the agent still works afterwards, you also need an eval suite and a pinned library version.</description></item><item><title>Neural Networks: Zero to Hero: Andrej Karpathy&apos;s eight lectures, from 150 lines of code to GPT</title><link>https://fdetimes.net/en/books-courses/karpathy-neural-networks-zero-to-hero/</link><guid>https://fdetimes.net/en/books-courses/karpathy-neural-networks-zero-to-hero/</guid><pubDate>Fri, 09 Oct 2026 02:54:18 GMT</pubDate><description>When a client asks why a model gave the wrong answer, an FDE who has written backprop and a tokenizer by hand can explain it. Someone who only knows how to call the API cannot.</description></item><item><title>ISO 20022 for FDEs: check the payment data before bringing AI into a bank</title><link>https://fdetimes.net/en/guides/iso-20022-payment-data-profiling-for-ai/</link><guid>https://fdetimes.net/en/guides/iso-20022-payment-data-profiling-for-ai/</guid><pubDate>Fri, 09 Oct 2026 02:00:57 GMT</pubDate><description>A bank that has moved to the new message standard does not necessarily have clean data. Measure how clean it is before a model learns the wrong things.</description></item><item><title>Extracting invoices and contracts with LLMs: schemas, business-rule checks and source citations</title><link>https://fdetimes.net/en/guides/llm-invoice-contract-extraction-validation/</link><guid>https://fdetimes.net/en/guides/llm-invoice-contract-extraction-validation/</guid><pubDate>Fri, 09 Oct 2026 02:00:51 GMT</pubDate><description>JSON that parses is not the same as JSON that is correct. This guide builds a pipeline that knows when to trust a figure and when to send it to a human reviewer.</description></item><item><title>Ten minutes with the client&apos;s leadership: turning pilot results into a signed decision</title><link>https://fdetimes.net/en/guides/present-pilot-results-to-executives/</link><guid>https://fdetimes.net/en/guides/present-pilot-results-to-executives/</guid><pubDate>Fri, 09 Oct 2026 02:00:41 GMT</pubDate><description>Your pilot may have run well. Spend the first minute of the meeting on architecture, though, and there may not be a second meeting.</description></item><item><title>The first FDE team: when to hire, what to measure and when to add an FDE Lead</title><link>https://fdetimes.net/en/analysis/first-fde-team-hiring-metrics-lead/</link><guid>https://fdetimes.net/en/analysis/first-fde-team-hiring-metrics-lead/</guid><pubDate>Fri, 09 Oct 2026 01:59:20 GMT</pubDate><description>Many people now treat FDEs as a cure-all. Phosai Labs, a services and consulting firm, offers a precise hiring signal and a 20% threshold for spotting when an FDE team has drifted into sales support.</description></item><item><title>Braintrust for FDEs: turning production failures into evals that block merges</title><link>https://fdetimes.net/en/tools/braintrust-evals-block-merge-regressions/</link><guid>https://fdetimes.net/en/tools/braintrust-evals-block-merge-regressions/</guid><pubDate>Fri, 09 Oct 2026 01:56:50 GMT</pubDate><description>At a customer site, you can only trust an answer to &quot;is the new prompt better?&quot; if a score runs on every pull request. Braintrust is built for exactly that job.</description></item><item><title>How to turn a finished deployment into a playbook, runbooks and a template repo for the next customer</title><link>https://fdetimes.net/en/guides/package-deployment-playbook-runbook-template-repo/</link><guid>https://fdetimes.net/en/guides/package-deployment-playbook-runbook-template-repo/</guid><pubDate>Fri, 09 Oct 2026 01:00:33 GMT</pubDate><description>What a team learns on a deployment usually leaves with the FDE when the project ends. This guide shows how to keep it in five documents and a template repo.</description></item><item><title>Multi-tenancy for AI systems: stopping cross-customer data leaks at four layers</title><link>https://fdetimes.net/en/guides/multi-tenant-ai-tenant-isolation-four-layers/</link><guid>https://fdetimes.net/en/guides/multi-tenant-ai-tenant-isolation-four-layers/</guid><pubDate>Fri, 09 Oct 2026 00:58:58 GMT</pubDate><description>When one RAG system serves many customers, a single missing WHERE clause can put one customer&apos;s documents into another customer&apos;s answer. To prevent that, the boundary has to be built into the database, the vector store and the encryption keys.</description></item><item><title>From FDE to founder: which habits carry over and which must be relearned</title><link>https://fdetimes.net/en/analysis/fde-to-founder-habits-that-transfer/</link><guid>https://fdetimes.net/en/analysis/fde-to-founder-habits-that-transfer/</guid><pubDate>Fri, 09 Oct 2026 00:57:46 GMT</pubDate><description>Y Combinator tells founders to work like engineers embedded with customers. The reverse is only half true, and the missing half decides whether you build a product company or a services business.</description></item><item><title>n8n for FDEs: build an LLM workflow in an afternoon, but know when to move to code</title><link>https://fdetimes.net/en/tools/n8n-llm-workflows-when-to-switch-to-code/</link><guid>https://fdetimes.net/en/tools/n8n-llm-workflows-when-to-switch-to-code/</guid><pubDate>Fri, 09 Oct 2026 00:56:25 GMT</pubDate><description>With drag-and-drop, you can have a customer demo running very quickly. Whether that demo can run in production depends on three things: the limits of the Code node, the scope of the licence and how the system scales.</description></item><item><title>Harvard&apos;s CS50 SQL: seven weeks on table design before you touch customer data</title><link>https://fdetimes.net/en/books-courses/harvard-cs50-sql-course-review/</link><guid>https://fdetimes.net/en/books-courses/harvard-cs50-sql-course-review/</guid><pubDate>Fri, 09 Oct 2026 00:54:15 GMT</pubDate><description>If you are about to read a client&apos;s database, redesign its tables and write safe queries, you can practise all three in a free course taught by David J. Malan and Carter Zenke.</description></item><item><title>Escaping the Build Trap: the book that teaches FDEs to measure their work by client outcomes, not feature counts</title><link>https://fdetimes.net/en/books-courses/escaping-the-build-trap-fde-review/</link><guid>https://fdetimes.net/en/books-courses/escaping-the-build-trap-fde-review/</guid><pubDate>Fri, 09 Oct 2026 00:54:09 GMT</pubDate><description>Melissa Perri wrote it for product managers, but the readers who need her 2018 book most are engineers who sit in a client&apos;s office and decide every week what to build next.</description></item><item><title>A2A and MCP: when your agent has to hand work to another vendor&apos;s agent inside a client&apos;s system</title><link>https://fdetimes.net/en/guides/a2a-vs-mcp-cross-vendor-agents/</link><guid>https://fdetimes.net/en/guides/a2a-vs-mcp-cross-vendor-agents/</guid><pubDate>Thu, 08 Oct 2026 23:57:43 GMT</pubDate><description>When the next link in the chain is an agent from a vendor whose code, memory and tools you cannot see, three things matter: use MCP for your own tools, use A2A to hand work to the other vendor, and expect the hard part to be authentication and tracing.</description></item><item><title>Andrew Ng&apos;s Machine Learning Specialization: three introductory courses, read through an FDE&apos;s eyes</title><link>https://fdetimes.net/en/books-courses/andrew-ng-machine-learning-specialization-fde/</link><guid>https://fdetimes.net/en/books-courses/andrew-ng-machine-learning-specialization-fde/</guid><pubDate>Thu, 08 Oct 2026 23:52:02 GMT</pubDate><description>Ten years on, the remake has moved to Python. For developers who want to become FDEs, the course&apos;s real value is that it teaches you a model well enough to explain it to a client.</description></item><item><title>HL7 v2, FHIR and Epic: integrating AI with hospital data through two pipelines</title><link>https://fdetimes.net/en/guides/hl7-v2-fhir-epic-ai-integration/</link><guid>https://fdetimes.net/en/guides/hl7-v2-fhir-epic-ai-integration/</guid><pubDate>Thu, 08 Oct 2026 22:57:14 GMT</pubDate><description>In a hospital, data reaches you by two routes: HL7 v2 pushes events to you, while FHIR lets you ask for state. If you pick the wrong route early on, your agent will either act too late or read the wrong data.</description></item><item><title>Peter Block&apos;s Flawless Consulting: the book for FDEs whose solution works but goes unused</title><link>https://fdetimes.net/en/books-courses/flawless-consulting-peter-block-fde/</link><guid>https://fdetimes.net/en/books-courses/flawless-consulting-peter-block-fde/</guid><pubDate>Thu, 08 Oct 2026 22:52:40 GMT</pubDate><description>On a client site, the hardest part is usually not the pipeline but the silent room after the demo. Peter Block wrote this book to teach you how to handle that room.</description></item><item><title>Practical Deep Learning for Coders: nine lessons from code to a working model</title><link>https://fdetimes.net/en/books-courses/fast-ai-practical-deep-learning-for-coders/</link><guid>https://fdetimes.net/en/books-courses/fast-ai-practical-deep-learning-for-coders/</guid><pubDate>Thu, 08 Oct 2026 22:50:43 GMT</pubDate><description>Jeremy Howard&apos;s free course has you running a model on real data in the first lesson. The theory comes later.</description></item><item><title>The Culture Map: hear American, German and Japanese client feedback correctly before it wrecks a deployment</title><link>https://fdetimes.net/en/books-courses/culture-map-erin-meyer-fde-feedback/</link><guid>https://fdetimes.net/en/books-courses/culture-map-erin-meyer-fde-feedback/</guid><pubDate>Thu, 08 Oct 2026 21:52:13 GMT</pubDate><description>American clients tend to praise first and criticise second, German clients say it straight, and Japanese clients convey a lot through what they leave unsaid. Erin Meyer&apos;s book helps FDEs hear all three correctly.</description></item><item><title>Building with the Claude API: Anthropic&apos;s free course covers half the engineering of the FDE job</title><link>https://fdetimes.net/en/books-courses/claude-api-course-fde-skills/</link><guid>https://fdetimes.net/en/books-courses/claude-api-course-fde-skills/</guid><pubDate>Thu, 08 Oct 2026 21:51:18 GMT</pubDate><description>Put the syllabus next to Anthropic&apos;s own FDE job posting and it becomes clear which modules deserve close study, which you can skim, and which parts of the work you will have to learn outside the course.</description></item><item><title>When the project sponsor quits mid-deployment: how an FDE keeps the work alive</title><link>https://fdetimes.net/en/guides/project-sponsor-leaves-mid-deployment/</link><guid>https://fdetimes.net/en/guides/project-sponsor-leaves-mid-deployment/</guid><pubDate>Thu, 08 Oct 2026 20:57:02 GMT</pubDate><description>The person who brought the project into the client has just resigned. The code still runs, but nobody is left to approve requests, sign off delivery or defend the project. The FDE is often the first to notice.</description></item><item><title>Accelerate: four metrics FDEs can use to measure software delivery speed and stability</title><link>https://fdetimes.net/en/books-courses/accelerate-dora-four-key-metrics/</link><guid>https://fdetimes.net/en/books-courses/accelerate-dora-four-key-metrics/</guid><pubDate>Thu, 08 Oct 2026 20:52:36 GMT</pubDate><description>First published in 2018, the book by Nicole Forsgren, Jez Humble and Gene Kim gives you hard numbers for the question clients ask most often: how fast and how reliably does our team get changes into production?</description></item></channel></rss>
