Bintou Diarra

What does it take to keep one of Europe’s biggest Rails monoliths running smoothly? In this week’s 'On Rails', Robby Russell chats with Florent Beaurain from Doctolib, where 400+ engineers work on a Rails monolith with 3M+ lines of code. Florent shares how they manage 84,000 tests, scale across 10 PostgreSQL databases, and keep Rails upgrades on track. Listen to the full episode here: https://lnkd.in/ez5U8BPs

117

2 Comments

🚀 Big news — I’m launching Le French Product, a podcast shining a light on French innovation in Product Management. And I couldn’t have asked for a better first guest: Robin Labrot (ex-lemlist, now founder of ListenUp! & SparkUp). In this episode, we tackle the questions every Product Manager is asking today: 👉 Is Product Management going to die because of AI? 👉 What’s the mindset shift from scaling a $25M ARR rocketship to starting again at $0 MRR? 👉 How can you run continuous discovery with minimal effort or better focus? Robin shares raw insights from bootstrapping at Lemlist, building AI-powered tools, and navigating France’s startup ecosystem. It’s a conversation about speed, focus, and the evolving role of Product Managers in the AI era. 🎧 Watch/Listen here: - Youtube : https://lnkd.in/emNuV5g7 - Spotify : https://lnkd.in/ea9ufwrC If you’re a Product Manager, founder, or just curious about how AI is reshaping our craft, this one’s for you. 👇 I’d love to hear in the comments: Do you think AI will replace parts of the PM role — or only make us stronger? #ProductManagement #AIProductManagement #ProductDiscovery #StartupGrowth #LeFrenchProduct

140

34 Comments

Big News: Stellar is launching a benchmark on Product development speed and quality. Here’s why I’m fully backing this initiative: One major weakness in many French companies: their execution speed. And speed matters for everyone: - Business - Finance - Marketing - HR But also for Product and Tech. And it’s even more critical because everyone in a tech company depends on Product. When a company ships fast, it: -Fixes bugs more quickly -Gains an edge over competitors -Collects feedback faster -Tests new ideas more rapidly Yet, more than 50% of French companies don’t deploy to production every week. And while there are exceptions, most companies should be doing it. This is a shared belief among all the CPOs at Stellar (co-founded by Timothe Frin from the Clef de voûte podcast). When Timothé and David created Stellar, they wanted to help leaders ship faster. So they came up with an idea: A benchmark on Product development speed – based on data from 110+ companies. The goal? To help Tech C-levels understand where their Product org stands. After 3 months of work, Stellar is launching its benchmark. Inside, you’ll find: -A system to measure Product development speed -Data from 110+ companies on speed and quality -Key indicators to assess Product quality -The 10 best-performing companies You are a C-level and want access to the benchmark? 👉 Get it here (100% free): https://lnkd.in/eXdTtp8x

57

3 Comments

Why do freelancers tell us so much about new tech trends? - With an average of 4 hours of training per week, they are pioneers in the adoption of new practices and technologies. - For their part, companies call on them when they haven't (or haven't yet) developed cutting edge expertise in-house. 𝗠𝗮𝗹𝘁’𝘀 𝗧𝗲𝗰𝗵 𝗧𝗿𝗲𝗻𝗱 𝟮𝟬𝟮𝟱 𝗿𝗲𝗽𝗼𝗿𝘁 leverages data from 200,000+ freelancers and millions of project searches on Malt. Trends captured by the report derive directly from thousands of companies using Malt to build real-world tech solutions. What do we learn?  ↗️ 𝗔𝗜 𝗶𝘀 𝗯𝗼𝗼𝗺𝗶𝗻𝗴 - 𝗼𝗳 𝗰𝗼𝘂𝗿𝘀𝗲: AI projects more than tripled in 2024 and demand for RAG even grew x17 in 2024! ↗️ 𝗟𝗼𝘄-𝗰𝗼𝗱𝗲/𝗻𝗼-𝗰𝗼𝗱𝗲 𝗶𝘀 𝗺𝗮𝗸𝗶𝗻𝗴 𝗮 𝘀𝘁𝗿𝗼𝗻𝗴 𝗰𝗼𝗺𝗲𝗯𝗮𝗰𝗸, fuel by synergies with AI: demand grew 40% in 2024. ↗️ 𝗦𝗼𝘃𝗲𝗿𝗲𝗶𝗴𝗻𝘁𝘆 𝗶𝘀 𝘁𝗼𝗽 𝗼𝗳 𝗺𝗶𝗻𝗱 𝗳𝗼𝗿 𝗰𝗼𝗺𝗽𝗮𝗻𝗶𝗲𝘀, who tend to adopt EU cloud and open LLMs. ❌ 𝗚𝗿𝗲𝗲𝗻 𝘁𝗲𝗰𝗵 𝘀𝗵𝗼𝘂𝗹𝗱 𝗯𝗲 𝗮 𝗽𝗿𝗶𝗼𝗿𝗶𝘁𝘆, 𝗯𝘂𝘁 𝗶𝘁 𝗶𝘀 𝗮 𝗯𝗹𝗶𝗻𝗱 𝘀𝗽𝗼𝘁: despite growing awareness of AI’s energy footprint, we haven’t seen a meaningful increase in demand for green tech skills. Download our full report: https://lnkd.in/eJi_TMwQ #AI #Cybersecurity #Cloud #Sovereignty #TechTrends #DigitalStrategy #LowCode #OpenSource #GreenTech #FreelanceTech #MaltTechTrends

49

🚀 No-code / low-code is back, and this time it's different! Our Malt Tech Trends Report shows: 230% rise in AI projects demand + 40% growth for no-code platforms. What changed? 🧠 GenAI shifted mindsets. Business teams building AI assistants suddenly realized: "If I can build this... why can't I automate the entire workflow?" At Malt 40+ employees across departments have become builders, creating AI assistants & automations that streamline meeting prep, support workflows, and more. 💡 The insight: The most powerful automation champions aren't in Tech: they're the people actually living with these processes every day. Give them tools and freedom, and they transform from task-followers into solution-creators. Want to learn more about our journey? Check out my latest article: https://lnkd.in/eZ3vNHk6

66

6 Comments

The Token Filter module enables you to embed these tokens straight into your content, functioning similarly to how shortcodes work in WordPress. Implementing Token Filter is straightforward: ➜ 1️⃣ Install the module using Composer: ➜➜ composer require drupal/token_filter ➜ 2️⃣ Edit a text format and enable the "Replaces global and entity tokens with their values" filter. ➜ 3️⃣ Add the "Token Browser" button to your toolbar to easily access available tokens. Once set up, you can embed tokens like [site:name] or [node:title] directly into your content. This feature allows for more dynamic and personalized content creation without custom code. #drupal #token ~~~ Quick note: I have a few consulting slots opening up for Drupal projects. Feel free to reach out for a quick chat about your Drupal projects! If you found this post helpful. ♻️ Please repost and follow. 📧 Join my newsletter; link in the first comment.

39

5 Comments

AI product development consists of the following steps : incubation -> experimentation -> scale -> operate. Do not let perfect be the enemy of the good in the first two stages, do not settle for anything less than perfect for the last two. What do you think ?

157

8 Comments

I’m thrilled to share that Benoit Reillier and I are working on a new edition of Platform Strategy this spring! 📚💪 While platform fundamentals remain the same, much has changed since the first edition released in April 2017. At the time, it was not always easy to convince leadership teams that platforms were such a powerful business model... today, we don’t need a long demonstration any more! Many platforms have grown exponentially. New regulations have reshaped the landscape. And emerging technologies—especially AI—are increasing redefining competition. Due to popular demand from our publishers, we’re updating the book with new insights, including: ✅ A fully documented Rocket Model 🚀 used by thousands of platforms around the world— allowing entrepreneurs to fast track their platform design, launch and scaling, and enabling practitioners to optimise established platforms. ✅ How to harness the power of AI—and avoid being “Chegg’d” when AI could be a substitute to your participants! ✅ How traditional organisations can combine platforms & AI with their existing business to create and orchestrate powerful ecosystems ✅ The increasing role of specific regulations, and how to navigate this new landscape ✅ How innovative platforms can be at the heart of the circular economy We'll keep you posted! #PlatformStrategy #AI #RocketModel #Marketplaces  

108

10 Comments

The real secret to stakeholder management? Translation. Not from English to French 🤣 but from technical to understandable. As Product Managers, we sit at the intersection of two worlds: 🧑‍💻 Engineers who speak in APIs, frameworks, and logic 📈 Stakeholders who speak in revenue, timelines, and business impact Bridging that gap is where real influence happens. Here’s how to translate technicality for proper stakeholder management 👇 ▶️ Simplify, don’t oversimplify - Break things down clearly without losing meaning. Clarity builds confidence. ⏩ Use visuals and analogies - A simple diagram or real-world example can make complex systems easy to grasp. ⏭️ Tailor your message - Your CTO and Marketing Lead don’t need the same depth. Speak their language. ⏭️ Be transparent - When stakeholders understand the why, alignment becomes natural. At its core, stakeholder management is about understanding, trust, and translation. Because great PMs don’t just build products, they translate impact. What’s your go-to strategy for explaining technical details to non-technical stakeholders? #productmanagement #stakeholdermanagement #leadership #communicationdkills #powergeng #growwithjoseph

30

23 Comments

I wanted to share some insights with you after migrating our Marketing Stream setup from SQS to Firehose 🙂 – in case you were undecided about which approach to take. - Cost Savings: Between the product cost itself and compute costs for processing, I am already seeing a savings of around 30%... - Flexible Output - Firehose allows a size and time buffer, which is great, since we can either shrink these down and have messages deliver to S3 in near real-time, or instead, can make the stream delay for up to 15 minutes and let the message size grow much larger. - Greater Efficiency - What prompted my Firehose experiment was that I found at times messages would arrive faster than I could pull/process, and delete them – AWS was responding with errors requiring messages to be deleted before more were pulled. Firehose has solved this. Although I could have further split the system into more queues per ProfileId, I thought trying Firehose would be a better solution. - Subscription Process - It is slightly different, in that, with SQS, a confirmation message is sent, which you would then make a request to finalize the subscription. With Firehose, this confirmation message is not part of the workflow. - Initial Setup - Even using the .YAML provided on github for CloudFormation, I still found the permissions issues to be tricky to get to work with the Firehose and S3 buckets… This was definitely the hardest part to get to work during my initial setup, and SQS never seemed to have this problem… FWIW – I have only done this with the Traffic and Conversion datasets – Stream Signals and Budget Usage is still utilizing SQS. Many people think of Marketing Stream as being synonymous with ‘Dayparting’, but it is so much more than just a means to get hourly data… Using Stream allows us to run a pipeline built on streamed data vs batched data. This considerably shrinks the pipeline’s footprint, since we only receive near real-time data that we need, as opposed to batching data, where the overwhelming majority of data is redundant with what our datastore contains. It makes the pipeline less expensive, more accurate throughout the day, easier to maintain, and more… For those interested, find the Overview here: https://lnkd.in/gRDT5Vch Check out this post, where I discuss some of the other benefits of Marketing Stream: https://lnkd.in/gtQacm4Q Amazon Ads Partners #btrmedia #retailmedia #adtech

34

23 Comments

"Teresa's approach felt like a toolkit to put a little bit of order into that messy process of discovery." 🎯 Discover how Sandrine Veillet led WebMD's journey to implement continuous discovery across a large organization! 🚀 This article explores: - The multi-phase approach to introducing continuous discovery 📊 - Strategies for training and coaching product teams 🏋️‍♀️ - Ways to involve stakeholders in the discovery process 🤝 - Challenges faced during implementation and adoption 🧗‍♀️ Key insights: - Start with raising awareness and getting buy-in from top management 📢 - Use templates and tools to make discovery concepts more tangible 🛠️ - Create a regular cadence for sharing and coaching discovery practices 🔄 - Involve stakeholders to foster trust and alignment 🤲 - Accept that behavioral change takes time and continuous effort 🕰️ Want to transform your organization's approach to product discovery? Check out the full article: https://buff.ly/4dgoXQY 📚 ❓What challenges have you faced in introducing or scaling continuous discovery practices in your organization? Share your experiences in the comments below.

27

2 Comments

🚨 New Data from Nomadic: We analyzed 143 million clicks and 1.4 billion impressions across some of our Tier 1 clients’ data in Google Discover — and the results are hard to ignore. 🔓 No-paywall content has: 📈 +90% more impressions on average ⚡ Nearly 2x more clicks than paywalled content But the takeaway isn’t “open everything.” ➡️ It means we need a strategic open-content layer in Discover — one that’s designed to funnel audiences into the paywalled experience. Think of it as a gateway, not a giveaway. Let’s stop thinking in binaries (open vs. closed) and start thinking in journeys. Discover brings attention — it’s our job to turn it into lasting value. I see publishers that close even the content that can be read in a non-subscription publisher just to get an e-mail or because every article counts in a metered paywall framework (you have a monthly counter that when you reach a certain amount of articles you got hit by paywall) Now is the time to: ✅ Define your open-content strategy ✅ Align editorial with Discover dynamics (entities, recirculation, CTR optimization) ✅ Build a funnel that connects visibility with subscriptions or monetization. ✅ If it's not a subscription-based content allow users coming from Google to always see the article that bring them to the site. Don't harm the user journey if the content is something that can be found in another publisher or if it can be easily replicated or answered by an LLM.

30

1 Comment

Browser Use 0.6.0 no longer runs on Playwright😮 We released a new version of Browser Use based directly on CDP (Chrome DevTools Protocol). Say hi to new era of web automation. New Browser Use is 🥷much less detectable (no Playwright leaks) ⚡faster 🎯more performant Technical blog explaining why we did this incoming soon👀 Go check it out https://lnkd.in/gj7fWsSd

340

14 Comments