Engineering Joe
ENGINEERING JOE / OBSERVE THE WORKSkip to the notebook ↗

01 / START WITH A QUESTION

WHO IS
ENGINEERING JOE?

Trace an answer. Find the failure. Fix the next run.

Follow the trace ↓
ILLUSTRATIVE RUN / NO LIVE TELEMETRYTRACE THE CLAIM. TEST THE FLOW.

02 / THE TRACE · ONE QUESTION, TWO RUNS

ONE QUESTION.
FOLLOW THE TRACE.

The first draft loses its citation. Follow the trace to find where.

INTRO-001 / RUN 1RUN 1 / CITATION NOT ATTACHED
04 / DRAFTJoe Dempsey - Innovation focused technologist based in London

First draft: no source citation attached. Continue to evaluation to check it.

Illustrative retrieval, using curated source:about. No model runs and no operation timings are simulated.

Continue to evaluation ↓

03 / EVALUATOR · ANALYZE, FIX, RERUN

FIND THE FAILURE.
FOLLOW THE FIX.

Run the checks, label the failure, then fix and rerun.

Three checks. One evidence contract.RUN 1 / READY TO EVALUATE
  • Source citation NOT RUN
  • Identity complete NOT RUN
  • Answer consistent NOT RUN
Review the supported introduction ↓

04 / ANSWER WITH EVIDENCE

MEET JOE.
SEE THE SOURCE.

A supported answer, with its source attached.

SOURCE PREVIEW / AWAITING YOUR REVIEW

Joe Dempsey - Innovation focused technologist based in London

S01 / Read Joe’s original About page ↗
EVIDENCE REVIEW / INTRODUCTION

Source citation · identity complete · answer consistent

Run the evaluation, then review its evidence.

Illustrative evidence review. No model is called in this walkthrough.

Explore the source connections
05 / THE NOTEBOOK IS OPEN

EVERYTHING IS
CONNECTED.

Explore the ideas, sources and writing behind this notebook.

LIVE CONNECTIONS / 22 ENTRIES

Choose your starting point.

PULL A THREAD
PersonTopicWritingTechnologyCurated relationships
Browse all connections as a list

The same subjects, writing and sources, available without the visual map.

Is DeFi Finally the Real Deal?

This article explores how DeFi has evolved from speculative beginnings to become a foundational, low-risk financial infrastructure. It highlights the shift towards safer protocols, growing institutional and regulatory support, and the merging of traditional and decentralised finance. Key developments include regulatory changes in the UK, the role of Ethereum as the Internet of Value, and new interoperability between banks and blockchains. The piece argues that DeFi is now delivering real-world impact and marks the start of a new era for global finance.

Open Is DeFi Finally the Real Deal?
LeetCode Style Interviews Are Over. It’s Not “Can You Code?” but “How Do You Build?

The technical interview is evolving. Instead of abstract LeetCode puzzles, companies now favour practical assessments that reflect real work. AI is central, candidates are expected to use tools like Copilot, demonstrating judgement and collaboration rather than memorisation. “How do you build?” not just “Can you code?”

Open LeetCode Style Interviews Are Over. It’s Not “Can You Code?” but “How Do You Build?
Unlocking the Potential of Gen AI Innovation within the Enterprise through Inner Source

Dive deep into web performance metrics, including tools and techniques for measuring and optimizing loading times. Discuss the significance of metrics like First Contentful Paint, Time to Interactive, and more.

Open Unlocking the Potential of Gen AI Innovation within the Enterprise through Inner Source
Optimise Your Research Workflow using GenAI, Part 2: Agents

Continuing the exploration of how Generative AI (GenAI) tools can enhance research workflows. In Part 1, we discussed foundational GenAI tools that streamline research processes like chatGPT & Perplexity. Here in Part 2, we focus on “agentic workflows” Combining tools and agents to automate complex workflows, and interact with large data sets and multiple sources, to enhance productivity through innovative capabilities.

Open Optimise Your Research Workflow using GenAI, Part 2: Agents
So are Deep Research Agents Enterprise ready?

The rapidly reshaping landscape of Deep Research agents this year has been incredible to the point where they look enterprise ready. Thoughts on benchmarks, APIs, and the real challenges for business use.

Open So are Deep Research Agents Enterprise ready?
Preparing Deep Research for enterprise

Exploring how to prepare Deep Research agents for enterprise use cases, focusing on three core ingredients: secure and reliable web search, robust orchestration architectures, and rigorous quality assurance (QA) pipelines. It compares leading search providers on enterprise readiness, discusses orchestration patterns (workflow, agentic, FSM), and details practical QA strategies to ensure trustworthy, auditable outputs at scale. The article provides actionable recommendations for building enterprise-grade research systems that balance flexibility, control, and trust.

Open Preparing Deep Research for enterprise
Skills Is All You Need.

Skills plus frontier models can now replace a surprising amount of bespoke workflow code. The catch is that the real complexity has moved into the harness: evals, traces, tool controls, and deployment boundaries.

Open Skills Is All You Need.
  • Explains · Curated relationshipEvaluation
    Evidence

    concepts relationship explicitly curated for this content; not a personal expertise claim.

  • Explains · Curated relationshipObservability
    Evidence

    concepts relationship explicitly curated for this content; not a personal expertise claim.

  • Written by · Curated relationshipJoe Dempsey
    Evidence

    Authorship explicitly declared in curated content metadata.

  • Mentions · Curated relationshipLangGraph
    Evidence

    technologies relationship explicitly curated for this content; not a personal expertise claim.

  • Mentions · Curated relationshipOpenTelemetry
    Evidence

    technologies relationship explicitly curated for this content; not a personal expertise claim.

  • Discusses · Curated relationshipAgentic systems
    Evidence

    topics relationship explicitly curated for this content; not a personal expertise claim.

  • Discusses · Curated relationshipAI / Generative AI
    Evidence

    topics relationship explicitly curated for this content; not a personal expertise claim.

  • Discusses · Curated relationshipAI reliability
    Evidence

    topics relationship explicitly curated for this content; not a personal expertise claim.

Joe Dempsey

Innovation focused technologist based in London

Open Joe Dempsey
Ethereum

LangGraph

OpenTelemetry

Agentic systems

Open Agentic systems
AI / Generative AI

Open AI / Generative AI
AI reliability

Blockchain

Open Blockchain
Deep research

Open Deep research
DeFi

Open DeFi
Engineering practice

Innovation

Platform engineering

Research workflows

Technical interviews

Open Technical interviews
ONE IDEA IS ONLY THE BEGINNING.Open the full knowledge explorer
06 / NOTES FROM THE FIELD

THINKING.
IN PRACTICE.

Research, experiments and lessons from building.

All writing
10 min read

Skills Is All You Need.

Skills plus frontier models can now replace a surprising amount of bespoke workflow code. The catch is that the real complexity has moved into the harness: evals, traces, tool controls, and deployment boundaries.

Agent HarnessAI EngineeringClaude Skills
Follow this thread in the map
5 min read

LeetCode Style Interviews Are Over. It’s Not “Can You Code?” but “How Do You Build?

The technical interview is evolving. Instead of abstract LeetCode puzzles, companies now favour practical assessments that reflect real work. AI is central, candidates are expected to use tools like Copilot, demonstrating judgement and collaboration rather than memorisation. “How do you build?” not just “Can you code?”

LeadershipInterviews
Follow this thread in the map
4 min read

Is DeFi Finally the Real Deal?

This article explores how DeFi has evolved from speculative beginnings to become a foundational, low-risk financial infrastructure. It highlights the shift towards safer protocols, growing institutional and regulatory support, and the merging of traditional and decentralised finance. Key developments include regulatory changes in the UK, the role of Ethereum as the Internet of Value, and new interoperability between banks and blockchains. The piece argues that DeFi is now delivering real-world impact and marks the start of a new era for global finance.

DeFIBlockchainEnterprise
Follow this thread in the map
26 min read

Preparing Deep Research for enterprise

Exploring how to prepare Deep Research agents for enterprise use cases, focusing on three core ingredients: secure and reliable web search, robust orchestration architectures, and rigorous quality assurance (QA) pipelines. It compares leading search providers on enterprise readiness, discusses orchestration patterns (workflow, agentic, FSM), and details practical QA strategies to ensure trustworthy, auditable outputs at scale. The article provides actionable recommendations for building enterprise-grade research systems that balance flexibility, control, and trust.

Deep ResearchGen AIEnterprise
Follow this thread in the map
SELECTED FROM THE NOTEBOOK
07 / THE WORK IN BETWEEN

CURIOSITY IS
A PRACTICE.

I'm Joe, a London-based technologist turning emerging ideas into working systems.

Between a research paper and a working prototype. Between a new technology and the problem it could solve. This is a notebook from that space.

More about my approach
  1. 01Identify.
  2. 02Test.
  3. 03Scale.
  4. 04Optimize.
THE NEXT CONNECTION

Keep the thought going.

New questions. New experiments. Every article, through RSS.

Follow the notebook