About
Passionate about empowering teams to revolutionize our clients’ businesses and have a…
Articles by Rachel
-
Why AI is relocating engineering work and what leaders need to know
Why AI is relocating engineering work and what leaders need to know
It's probably an understatement to say the pace of AI-driven change is intense at the moment. Just reading or watching…
31 Comments
Activity
11K followers
-
Rachel Laycock shared thisAre we teaching junior engineers to think, or just to prompt? I had a fantastic time debating this topic at LeadDev's LDX3 New York alongside Nivia Henry and Mattan Bitner, moderated by Randy Shoup. I’ve been thinking a lot about the concept of cognitive surrender - the tendency to accept AI-generated outputs without fully understanding the reasoning behind them. As AI accelerates software development, our ability to generate code is increasingly outpacing our understanding of design and architecture. Historically, engineers developed that judgment through experience, mentorship, pair programming, and working through problems the hard way. That's why one of the best pieces of advice for junior engineers that emerged from the discussion was: don't prompt, don't write code, first talk about how you would solve the problem first. The real challenge isn't generating more code. It's developing the judgment to understand systems, make sound architectural decisions, and know when AI is helping versus when it's leading us astray. As code volumes continue to grow, that same shift applies to code reviews. Rather than reviewing every line, engineering teams should spend more time reviewing design decisions, architecture, and system impacts. The future of engineering isn't just about better prompts. It's about better judgment. #LDX3 #LDX3NewYork
-
Rachel Laycock shared thisWe made it to Recife, Brazil! It’s time to put together the next edition of the Thoughtworks Technology Radar, and it has never been more interesting to bring a group of technologists together to debate what we are seeing. In a world where it has never been easier to build a tool, launch a product or create convincing marketing around it, the signal can quickly get lost in the noise. That is why the Radar matters. Experienced technologists comparing notes, challenging each other and drawing on what they have seen in real-world delivery to offer practical, independent advice. I’m looking forward to the debates.
-
Rachel Laycock shared thisLooking forward to being at LDX3 by LeadDev in New York next week. I’ll be joining two conversations on questions I’ve been thinking about a lot lately: are we teaching junior engineers to think, or just prompt? And what might our workforce look like on the other side of AI? If you’re at LDX3, come along and join the conversation. I’d love to hear your perspectives too. Check out my sessions and the full agenda here: https://lnkd.in/em-97dNS #LDX3NewYork
-
Rachel Laycock shared thisI’m grateful to be nominated for the 2026 World Leaders in Data and AI [WLDA] Technology Leadership in AI Award. For me, this recognition is as much about the people as it is the technology. Meaningful AI transformation doesn’t happen in isolation, it comes from teams bringing together different perspectives, challenging assumptions and working through the hard questions of how we use AI responsibly and effectively. I’m fortunate to work alongside some brilliant colleagues who do exactly that every day. Thank you to everyone who has been part of that journey.
-
Rachel Laycock shared thisAI is producing more code than humans can realistically review, but I’m increasingly convinced the answer isn’t faster code review, it’s less code review. I recently had a great discussion with Brian Houck from DX about what we lose if code review can’t keep up. Knowledge sharing, mentoring, architectural alignment, collective ownership. All important. My question is: why are we waiting until code review to do any of them? Pair earlier. Design together. Automate what can be automated. Encode architectural constraints. And use human review where human judgment is actually valuable, rather than requiring someone to inspect every change because that’s how we’ve historically created confidence. Did AI break code review? Or did it expose how much we’ve been asking it to do? My latest Rachel’s Ramblings: https://lnkd.in/e6pqivMN
-
Rachel Laycock shared thisI've really enjoyed watching our Global Innovation Hackathon unfold over the past few weeks. Seeing people from across Thoughtworks come together, challenge each other's thinking, experiment with AI and turn ideas into things that could solve real problems is exactly the kind of curiosity I want us to keep creating space for. Choosing which ideas would make it through to the final was genuinely difficult. There were so many strong ones. But perhaps the bit I've enjoyed most is seeing that, for some teams, not making the final hasn't meant the idea stops. Some are already finding their way into teams, projects and client conversations. To me, that's the real measure of whether something like this worked. Not just who wins, but what we learned, what we tried, and what keeps going long after the hackathon is over. A huge congratulations to everyone who took part, and good luck to our finalists. I'm looking forward to seeing where you take these ideas next.
-
Rachel Laycock shared thisFor the last few years we have been constantly innovating in how we build, modernize and run software. This culminated in the first release of AI/works™ in January this year. Like any early release it was far from perfect and instead of just adding more features we have spent the last few months looking at what underlying rigor is required to take this to scale. That's the thinking for the latest release where we've strengthened governance, security and observability because moving from AI experimentation to enterprise delivery requires more than speed. It requires confidence in how these systems operate, how they're controlled and how they perform. For me, that's the standard we should be aiming for: innovation with the engineering rigor needed to stand behind what we build. Take a look at what's new in AI/works™: https://lnkd.in/eJrKWsKa
-
Rachel Laycock shared thisAs we move further into agentic systems, I think we're going to have to get much more sophisticated about how we think about models. Not every task an agent performs needs frontier-level intelligence. Some steps need deep reasoning. Others are high-volume, repetitive or very specific to your business. Using the biggest, most capable model for every one of those steps isn't necessarily the best architecture, and the economics certainly won't make sense at scale. Our teams have been experimenting with NVIDIA Nemotron 3.5 Lightning to understand what a more specialized model approach could look like. What I found interesting was how quickly they could adapt the model to a particular domain, change how it behaves without meaningfully degrading its broader capabilities, and significantly improve inference speed. For me, the bigger story isn't really about one model. It's that agentic AI is increasingly going to be a system of models, and we're going to need to make deliberate choices about which model does which job. That changes the enterprise AI question from "which is the best model?" to something much more interesting: where do we actually need frontier intelligence, and where would greater control, specialization and better economics produce a better system? Gustavo A. Lujan, Allen Roush and Andy Nolan share what they learned, and all the technical detail, here: https://lnkd.in/e5qFhqFN
-
Rachel Laycock shared thisFor my third post on Rachel’s Ramblings, I’ve been pondering the growing gap between non-technical people who can now build surprisingly full-featured apps with AI, and their understandable question: why isn’t the engineering team suddenly 10x faster, and why can’t we just put this thing into production? I think the gap is trust. Production-grade software requires more than working code. It requires good engineering judgement about security, resilience, operability, scale and all the things that determine whether we can actually trust a system. That’s what led me to: Citizens build. Agents execute. Experts govern. Not as three new job descriptions, but as a shift in where value sits, and a way of thinking about what it will really take to enable anyone to turn an idea into production-grade software. I'd love to hear others thoughts. https://lnkd.in/dXcZ5tMN
Rachel Laycock replied to a comment
2w
Wolf Schlegel - Quite possibly! As code volume rises, reviewing the tests and evidence may tell us much more than trying to read every line the agent produced.
Rachel Laycock replied to a comment
2w
Parthasarathi Bhattacharjee - I agree. Pairing still has so much value. As AI takes on more of the implementation, the conversations and opportunities to learn from each other may become even more important.
Rachel Laycock replied to a comment
2w
Jaeseok Park - Exactly. It’s less about preserving the process and more about preserving the outcomes we needed from it, then finding better ways to achieve them as software development changes.
-
Rachel Laycock liked thisRachel Laycock liked thisI was so lucky that my good friend and research collaborator Brian Houck could visit us at #UVIC yesterday (pictured with our new colleague Amber Horvath). Brian delivered an **incredible** talk! Through a really engaging story (at Brian speed :) ), he brought together findings from over 12 studies to share what he has learned about the impact of AI on software development over the past few years. A couple of takeaways (sorry we didn't record it but I'm sure this is a talk he will give again): "AI is coming for our deep work". And as developers manage multiple agents, it is like we are playing "multiple games of chess at the same time" when the skill we spent years learning is to play a single game. Brian also talked about why context is a first class engineering artifact, referring to the CAFE(S) diagnostic framework for assessing context quality (work we just published at ACM Queue this past week!). He also gave a shout out to Annie Vella for identifying the productivity-experience paradox! He concluded the talk by reminding us we have to keep listening to our developers!
-
Rachel Laycock liked thisRachel Laycock liked thisMy first time seeing my new book, Building Resilient Distributed Systems in the flesh! Thanks to Barry O'Reilly for loaning me his newly purchased copy for the photo at #GotoCPH2026
-
Rachel Laycock liked thisRachel Laycock liked thisLook at what arrived in the mail today... thanks Francisco Trindade! 🥳 This is a great contribution to the body of engineering management knowledge, blending some of my favourite topics like systems thinking, leading and continuous improvement. This will be a very useful book for many Engineering Managers (and others too I think!) I was able to do an early review of this book. Well done again Francisco Trindade and I'm excited to read the final copy. I'll add the link in the comments below.
-
Rachel Laycock liked thisSure, you can come along for my thoughts on how to handle uncertainty with AI (and robots!), but the main draw is going to be some awesome insights from the other speakers including some grade A snark from Corey Quinn! O11yday London, 7th of October.Rachel Laycock liked thisHoneycomb's coming to London! For this upcoming Observability Day, we're hosting the inimitable Sam Newman, keynoting on "Cobots, dark factories and architecting for AI," as well as hearing from our friends Corey Quinn of Duckbill Group on the cost structures behind AI workloads, Yana Chen of Baseten on ensuring reliability for ML inference at scale, and Brian Scanlan and Inokentii M. of Fin.ai on their journey and learnings from scaling engineering productivity. It should be a banger of an event 🔥 Excited to see you all there! Register here: https://go.hny.co/4gGufKn
-
Rachel Laycock liked thisRachel Laycock liked thisAnd 15.5 years later, I’m closing a long and special chapter. September marks my departure from Thoughtworks. I traveled the world and lived in four countries besides home. For a while, everything I owned fit into a single suitcase. The first picture is from the end of 2015 when I moved apartments in Chile and I was surprised by the two extra bags (it turns out a skateboard takes space). I met so many smart, kind and cool people. Scrolling through my pictures, so many memories came flooding back: places, projects, teams, clients, late nights, celebrations, and all the unexpected moments in between. There have been a constant and fair amount barbecues too... you are welcome, I enjoyed every single one of them. The second picture is from a very memorable barbecue in Porto Alegre in 2013 with a few TWers, and the third is my dad getting the fire and feast ready (and the costelão) for a few clients and colleagues at his place in Guaíba (2015). Thank you to all my Clients and TWers who made it such an incredible ride. I’m very grateful for everything I learned, the opportunities I was given, and the people I crossed paths with. On to the next chapter.
-
Rachel Laycock liked thisAfter 19 years in the company, I mustered the courage to propose content for Martin Fowler’s website. That content is now live. It’s a series that I’m calling “Never Send the Slides”. It’s a provocative name, I hope, because it calls out one of the biggest antipatterns that impedes effective, engaging, “conspicuously human” presentations. The goal? To articulate what it might take to deliver effective presentations in a knowledge work context. I must 🙏🏾 thank Santosh Mahale, Lubaina Rangwala, Raghvi Behl, Jade Daubney-Bennett, Gill Taylor, Gayathri Moghe and Anuroop Krishnan for reviewing the content at various stages and cheering me on. ICYI here are the first two pieces. 1️⃣ Introduction: https://lnkd.in/gPTQUsZR 2️⃣ Do you even need a presentation?: https://lnkd.in/g96Kz4Kp If you like what you see, please watch this space 👀. I'm writing almost every week, though the pace of publication may be a bit slower. And if you have bouquets, brickbats, criticism, encouragement, love or hugs, send them my way 🙂.Rachel Laycock liked thisLike me, Sumeet Gayathri Moghe is tired of poor presentations with bad slide decks. He's started to write a series of posts on how to avoid these calamities, beginning with a post that questions whether a presentation is needed at all. https://lnkd.in/gSmQ4yv4
Experience & Education
Projects
-
Technovation Challenge 2014
- Present
Coached two teams of students to develop apps for Technovation Challenge, an international technology entrepreneurship program for girls. Both teams advanced to the final round, Technovation Challenge World Pitch held at Intel HQ in Santa Clara, CA. FieldTripper, The Middle School team, won 1st place out of 150 and $5,000. SafeTeen, the high school team, placed in Top 8 out of 225.
Other creatorsSee project -
Personal Finance and Workplace Investment Firm
-
Lead analyst on this project from Assessment to Inception to Delivery.
Worked with team to deliver a digital experience targeted toward a new customer demographic and coach client on agile practices.
Created a bridge between the business, design and development teams, ensuring that the correct value and vision were delivered.
Aided in creating a productive environment by taking lead on team-building and wellness activities and events.Other creators
Recommendations received
1 person has recommended Rachel
Join now to viewOther similar profiles
Explore more posts
Explore top content on LinkedIn
Find curated posts and insights for relevant topics all in one place.
View top contentOthers named Rachel Laycock
20 others named Rachel Laycock are on LinkedIn
See others named Rachel Laycock