Building LLM Applications is now a top 10 course on Maven (20% off ends Sunday)
And What to Build Instead of AI Agents
Welcome to Vanishing Gradients!
This issue kicks off with some news: Building LLM Applications was just named a top AI course in the Maven 100, and it’s now ranked #9 overall, #7 in AI, and #3 in Data & Engineering!
To mark the occasion, Maven’s offering 20% off until Sunday, and our scholarship applications close the same day.
Alongside that, we’ve got:
A new podcast with Hamel Husain on why most teams get evaluation wrong
A guest post on five reliable alternatives to agents
Stefan Krawczyk on the two iteration loops every LLM app needs: one for development, one for production
Highlights from our Berlin meetups with Rasa, spaCy, and Native Instruments
And a conversation with Alto Pharmacy’s Sudarshan Seshadri on building AI systems you can actually trust
Quick links below to what’s coming up in my data/AI life, what just dropped, and how to plug in:
📺 Live Online Events
📩 Can’t make it? Register anyway and we’ll send the recordings.
→ June 20 — 10x Your Productivity by Building Personal Agents with MCP with Skylar Payne (AI executive for startups. Ex-Google. Ex-LinkedIn.)
→ June 30 — Workshop: From Images to Agents with Ravin Kumar (DeepMind)
→ July 1 — Lightning Lesson: GenAI’s 4 Pillars with John Berryman (AI Consultant and Builder, Co-Author of Prompt Engineering for LLMs, ex-Github)
→ July 2 — Human-Seeded Evals (Live Podcast) with Samuel Colvin (Pydantic, Logfire)
🎙 Podcasts & Recordings
→ A Field Guide to Rapidly Improving AI Products – Hamel Husain
→ High-Stakes AI Systems and the Cost of Getting It Wrong – Sudarshan Seshadri
📍 In-Person Events
→ June 24–25 — VentureBeat Transform (San Francisco; use the code 25VanishingGradients for 25% off!)
→ July 7–11 — SciPy 2025 (Tacoma, WA)
🎓 Courses
→ June 22 — Scholarship deadline for the July cohort of Building LLM Applications
🎓 Named a Top AI Course in Maven 100 — 20% Off + Scholarships Close Sunday
Big news: Building LLM Applications for Data Scientists and Software Engineers, the course I co-teach with Stefan Krawczyk (Agentforce @Salesforce), was just selected as a top course in AI, Data, and Engineering on Maven.
It’s also climbed the rankings—now sitting at #9 overall, #7 in AI, and #3 in Data & Engineering.
This recognition comes after two cohorts and 150+ learners from companies like Netflix, Meta, Adobe, Amazon, and TikTok have built production-grade LLM apps with us.
The next cohort runs July 8 – August 2, timed for the Americas and Europe/Africa. We’ll go deep on:
Designing and deploying scalable LLM applications
Debugging, monitoring, and building reliable AI systems
Evaluating and building agentic workflows
Shipping portfolio-worthy multimodal projects
🧰 You’ll also get:
$500 in Modal credits (thanks to Charles Frye)
$300 in Google Cloud & Gemini credits for the first 100 registrants (thanks to Ravin Kumar and Paige Bailey)
6 months of free access to Prodigy from Explosion (thanks to Ines Montani)
More credits to be announced soon
To celebrate the course’s selection in the Maven 100, Maven is offering 20% off until Sunday.
👉 Sign up here
🧪 Also: one of our Lightning Lessons—Mastering LLM Application Testing by Stefan Krawczyk—was just named one of Maven’s top 100 Lightning Lessons. It walks through there two iteration loops every LLM app needs (one for dev, one for prod) and has already been watched by nearly 700 builders.
👉 Watch it here
We’re also offering a limited number of fully funded scholarships for this cohort. Applications are open through Sunday, June 22, and we’ll announce the three recipients the following week.
👉 Apply here
✍️ What to Build Instead of AI Agents
I recently co-authored a piece with Jordan Cutler for his High Growth Engineer newsletter, breaking down when agents fall short and the patterns that often work better in production.
We walk through five practical alternatives:
Prompt chaining
Parallelization
Routing
Orchestrator–worker setups
Evaluation loops
These patterns build on Anthropic’s framework for effective agents, with examples of how to apply them in real-world systems.
📖 Read the post: Stop Building AI Agents
🎙️ New Vanishing Gradients Podcast: Hamel Husain on Getting Evals Right
Most teams get AI evaluation wrong. That’s not a hot take—it’s a hard truth we’ve seen firsthand in our course and in production systems.
For more context, we’ve started releasing some of the standout guest lectures from our Building LLM Applications course. This episode features Hamel Husain (ex-Airbnb, ex-GitHub, now independent consultant), who joined us to break down why treating evals as “just a metric” misses the point entirely.
“If evals is just a metric, then you're thinking about evals wrong. It’s not a metric—it’s an entire process.”
We talk about:
Why most teams struggle to define what “better” even means
How to discover what you (and your users) actually want
What it really takes to improve AI systems in the wild
🎧 Listen to the full conversation:
→ Spotify
→ Apple
→ YouTube
→ Show notes and all episodes
🎧 High Signal: Trustworthy AI in High-Stakes Domains with Sudarshan Seshadri
“As data leaders become the backbone of decision-making, one of the most important skills is building real muscle around making big, irreversible decisions.”
In this new episode of High Signal, Sudarshan Seshadri (VP of AI, Data Science, and Foundations Engineering at Alto Pharmacy) shares what it takes to build high-stakes AI systems that actually earn trust.
We cover:
Supporting human judgment in safety-critical environments
Using LLMs for structure, not generation
Building feedback loops into real-world ML systems
Why optimizing for speed can break trust in healthcare
📍 Berlin Meetups: Real-World AI, Agents & Evals
Thanks to everyone who joined us in Berlin this week for two packed evenings of real-world LLM strategies, lightning talks, and open discussion.
At Native Instruments, we explored PDF processing, eval-driven development, and building AI products without the fluff, with great talks from Ines Montani, Matthew Honnibal, Vincent Koc, Shiva Banasaz Nouri, Fred O’Loughlin, Harrison Pim, and more.
At the Agents & Evals meetup (co-hosted with Rasa), we kept it informal—walk-on lightning talks, high-signal conversations, and plenty of nerdery under the Berlin sun.
🙏 Big thanks to our hosts at Native Instruments and Rasa for making it happen.
If you want to know a bit more about them, check out the Native Instruments recap and the Rasa meetup reflections.
📸 A couple of photos below (see the posts above for more!):
Want to Support Vanishing Gradients?
If you’ve been enjoying Vanishing Gradients and want to support my work, here are a few ways to do so:
🧑🏫 Join (or share) my AI course – I’m excited to be teaching Building LLM Applications for Data Scientists and Software Engineers again—this time with sessions scheduled for both Europe and the US. If you or your team are working with LLMs and want to get hands-on, I’d love to have you. And if you know someone who might benefit, sharing it really helps.
📣 Spread the word – If you find this newsletter valuable, share it with a friend, colleague, or your team. More thoughtful readers = better conversations.
📅 Stay in the loop – Subscribe to the Vanishing Gradients calendar on lu.ma to get notified about livestreams, workshops, and events.
▶️ Subscribe to the YouTube channel – Get full episodes, livestreams, and AI deep dives. Subscribe here.
💡 Work with me – I help teams navigate AI, data, and ML strategy. If your company needs guidance, feel free to reach out by hitting reply.
Thanks for reading Vanishing Gradients!
If you’re enjoying it, consider sharing it, dropping a comment, or giving it a like—it helps more people find it.
Until next time ✌️
Hugo