Sloan the DEV Moderator

Posted on Oct 14

Welcome Thread - v 397

#welcome

Leave a comment below to introduce yourself! You can talk about what brought you here, what you're learning, or just a fun fact about
yourself.
Reply to someone's comment, either with a question or just a hello. 👋

Come back next week to greet our new members so you can one day earn our Warm Welcome Badge!

Top comments (291)

FrancisTRᴅᴇᴠ • Oct 14

Hey all. Welcome to DEV and thank you for choosing this platform!!

As mention on my last week thread message and the week before that, I am waiting on introducing myself to the welcome thread (properly) @javz 3 more weeks till Welcome Thread v400!

In other case, what brings you on DEV? Any good projects you are working on?

UK • Oct 17

Hello!

Ouroboros Collective (T. Markgraf) • Oct 15

Hallo 🫠

Leonardo Weber • Oct 16

Hello

Genadi Avetisov • Oct 17

Hey Francis, thanks for keeping the welcome thread going! 👋

I'm working across software engineering, AI systems, and developer tooling at Ghost Labs, and lately I've been exploring Rust. What kinds of Rust projects or topics have sparked interesting discussions here recently?

Achmad Hendarsyah • Oct 17

Hello! I'm here to learn, share what I discover, and connect with fellow developers! Currently exploring new ideas and building projects to sharpen my skills. Looking forward to learning from the community and sharing my journey along the way!

Aneesh Hebbar • Oct 17

Hello!

Gleb • Oct 16

Hello!

Olamiji Michael Segun • Oct 17

Hello 👋

Loi Bui • Oct 16

Riddhi Rawool • Oct 16

hello

Martin Krause • Oct 14

Hi everyone!

Welcome to the dev.to-community!

I'm Martin, and I worked as a Senior Front-End Architect and Engineering Manager at some of the world's largest digital agencies, where I led front-end teams on international projects. After 15 years, I left.

Now I travel the world as a professional scuba diver and work as a contracted Tech Lead, Solution Architect, and technical Recruiter for international clients. Additionally, I write about technology on dev.to and Golem (German).

A life, Jimmy, you know what that is? It's the shit that happens while you're waiting for moments that never come.

That is my advice for everyone. Jump right in. Start writing! Today! That's how I became a published author for No Stach Press and Packt Publishing - who knows what's waiting for you?

My latest book Claude Code Beyond the Prompt will be available in November. Over the next few weeks, I'll run a series on dev. to revealing the chapters one-by-one. Take a look :)

If I have enough time, I write more in-depth articles and thought pieces here on dev.to, for example: Toward Agent Native Development.

My first book, The Complete Developer, is an easy, hands-on introduction to building web applications with Next.js, MongoDB, and Mongoose. It is part of multiple university syllabi in the USA and India - you can get it from No Starch Press

I recommend following the general tags webdev, ai , career and productivity and javascript, react, CSS. Head over to your Dashboard and adjust the tag weight to personalize your home feed.

Here's a list of articles to start your journey as a successful developer and a great writer:

If you are looking for your next job, I highly recommend buying the Book De-Coding the Technical Interview Process by Emma Bostian. It is hands down the best preparation for every technical interview.

If you're interested in my articles, have questions, or want to get in touch, don't hesitate to contact me on dev.to/martinkr

And remember to start posting today!

Cheers!

𝗝𝗼𝗵𝗻 • Oct 14

Hey Martin! 👋

Transitioning from 15 years in agency leadership to balancing freelance tech lead work with professional scuba diving is such a cool journey. Talk about mastering work-life balance!

I really appreciate how consistently you encourage newcomers to start writing and sharing their journey early. With your experience mentoring and authoring The Complete Developer, what’s the single biggest barrier you see developers face when shifting from just consuming content to actually publishing their first post or open-source contribution?

Great having your perspective in the thread! 🚀

Martin Krause • Oct 14

Thank you for asking. The biggest single barrier "is not doing it". You will only improve a skill by using it. Write a ticket, raise a PR, publish a Library - don't hesitate; do it. Do it yourself, not ask an LLM, and you will learn something every time.

𝗝𝗼𝗵𝗻 • Oct 14

Couldn't agree more, Martin. The friction of wrestling with an issue and figuring it out yourself is where the actual muscle memory gets built. Shortcuts just don't teach you how to think through the messy edge cases.

Really appreciate the grounded perspective and the push to just build!

Cyrille Labesse • Oct 15

Ola Martin!

Loving the work-life balance piece, it's great to see more and more of this, hallelujah!

I’m Cyrille 🇫🇷 a developer for 20 years, mostly RoR. Interested in our relationship to machines, the good and the bad 😊

I’m here to read, swap stories, and hear how others are handling these wild changes!

Ali Atif • Oct 16

Rachit Dubey • Oct 16

hello

Ouroboros Collective (T. Markgraf) • Oct 15

Komme auch aus Deutschland:)

Victor Cervantes • Oct 16

Wow, a lot of recommendations, thanks for your help.
I hope to learn a lot from the people in this community, have a great day.

Thomas Bonnet • Oct 14

Hello and welcome to dev.to Community ! 👋🏼😉

Muhammad Hamza • Oct 15

Hi everyone! I'm Hamza. I've been building small web tools with Next.js, and lately I've been working with MediaPipe FaceMesh to measure facial proportions directly in the browser, without uploading the photo anywhere. Looking forward to learning from this community and sharing what I've figured out along the way.

𝗝𝗼𝗵𝗻 • Oct 15

Welcome to DEV, Hamza! 👋

Running MediaPipe FaceMesh purely client-side is a huge win for user privacy. People definitely appreciate biometric data staying on their own device.

How has the landmark detection performance held up across mobile browsers - did you run into any thermal throttling or frame-rate drops?

Muhammad Hamza • Oct 16

Thanks, John! Good question. In my case it turned out to be easier than I expected, because the tool works on a single photo rather than a live camera feed. It runs FaceLandmarker in IMAGE mode, detects once and stops, so there is no frame loop running long enough for thermal throttling or frame drops to show up.

The costs I did hit were elsewhere:

First load. The WASM runtime and the model only download after someone picks a photo (a dynamic import), so they don't slow down the page itself. I create the landmarker once and reuse it, so a second photo in the same session is much faster.
GPU delegate. It fails on some mobile browsers, so I wrap it in a try/catch and fall back to the CPU delegate. For one still image, CPU is perfectly fine.
Photo quality matters more than speed. Most bad results on phones came from selfies taken too close or with the head turned, not from performance.

If I ever add live camera tracking, throttling would become a real concern, and I'd probably cap the frame rate and run detection on every other frame. Have you seen good results with that approach?

𝗝𝗼𝗵𝗻 • Oct 16

That lazy-loading setup with dynamic imports and the CPU fallback is really smart. Keeping the initial bundle lean and handling inconsistent mobile WebGL/GPU delegates via try/catch saves so many headaches.

Regarding live tracking: yes, running detection on every 2nd or 3rd frame (or capping inference at 10–15 FPS) is the standard and works extremely well. In practice, you can let the camera feed render smoothly at 60 FPS while decoupling the inference loop using requestVideoFrameCallback or an OffscreenCanvas inside a Web Worker. For facial proportion tracking, users won't notice the lower inference rate, but their phones will definitely thank you for avoiding thermal throttling.

Really appreciate you sharing the breakdown of the real-world constraints. Great engineering decisions!

Muhammad Hamza • Oct 16

Thanks, John, that's really useful. Decoupling the camera render from the inference loop with requestVideoFrameCallback is exactly the piece I was missing, and moving detection into a Web Worker with OffscreenCanvas would keep the main thread free for the UI too.

For proportion measurements specifically, a lower inference rate might even help: averaging the landmarks over a few frames should smooth out the small jitter that a single frame has, which matters when you're measuring differences of a fraction of a percent. Something to try if I build a live mode. Appreciate you sharing this!

Ojaswin Aggarwal • Oct 15

Hey everyone! I'm Ojaswin, a software engineer who enjoys building things, solving engineering problems, and exploring how technology works under the hood.

I'm particularly interested in AI engineering, LLM-powered applications, backend development, and distributed systems. I enjoy experimenting with new technologies, building practical projects, and learning by getting my hands dirty with code.

On DEV, I plan to share my learning journey, technical discoveries, project breakdowns, debugging experiences, and lessons from building real-world applications. I believe the best way to learn is to build, question, experiment, and share what you discover along the way.

I'm also looking forward to connecting with fellow developers, learning from the community, and contributing wherever I can.

🔗 GitHub: github.com/Ojaswin30

Looking forward to learning and building with you all. Let's connect!

𝗝𝗼𝗵𝗻 • Oct 15

Welcome to the DEV community, Ojaswin! 👋

The intersection of distributed systems and LLM engineering is such a fascinating space right now - especially when it comes to tackling latency, state management, and actual system reliability under load.

Your plan to share debugging stories and project breakdowns is spot on; some of the absolute best reads on DEV are candid post-mortems showing what broke, why it broke, and how you fixed it.

What kind of backend or AI architecture are you experimenting with at the moment?

Alessio Bianchini • Oct 14

Hi, I'm Alessio, software dev from Italy. Joined for the Hacktoberfest AI challenge and just wrote up a weekend project—built a small local receipt parser with Ollama to stop manually typing invoice data into spreadsheets: ledgerlocal

𝗝𝗼𝗵𝗻 • Oct 14

Welcome to DEV, Alessio! 👋

Always great to see folks building practical, offline-first tooling around local SLMs and Ollama rather than defaulting to cloud APIs.

What’s been your biggest takeaway so far regarding the latency vs. precision trade-off when running models directly on client hardware?

Looking forward to your posts around here! 🚀

Alessio Bianchini • Oct 15

Thanks! Honestly, 3B models like phi4-mini are plenty fast (2-3s per doc), but they get tripped up by layout quirks—like mistaking the customer for the vendor on Italian receipts.

Instead of moving to a heavier model that kills battery and takes 20s, I found it way better to stick with the small model and just patch the edge cases with regex and a quick edit UI. Fast SLM + deterministic code in front of it seems to be the sweet spot.

𝗝𝗼𝗵𝗻 • Oct 15

That hybrid approach is spot on and so underrated.

Pairing a lightweight SLM with deterministic post-processing and a quick UI beats throwing brute compute at the problem every time, especially on client hardware where thermal throttling and battery life actually dictate UX.

Getting 95% of the extraction done in 2 seconds with deterministic guardrails is the exact kind of pragmatic engineering more teams need to adopt. Would love to read a full case study or architectural breakdown on that if you end up writing one!

Ali Atif • Oct 16

Hi everyone!

I’m Ali, a final-year CSE (AI) student from India, and a fresher looking to start my career as a Software engineer.

I work mostly with React, Next.js, Node.js and FastAPI, and I’ve been interning as an AI engineer for the past few months. On the AI side, I’ve built RAG pipelines (ChromaDB + BM25), multi-agent systems with LangGraph, and integrated LLMs like Gemini and Groq LLaMA into real products. My favorite win so far was cutting an API’s latency by 81% just by fixing an N+1 query problem.

I’m here to learn from people who build in production, share what I figure out along the way, and hopefully write my first post soon. If you’re working on anything with LLMs or backend performance or anything else, I’d love to hear about it.

What’s the most useful thing you learned?

Emmanuel Gyamfi • Oct 14

Hi everyone!
I’m new here and wanted to drop in and introduce myself. I'm currently focusing my learning on Artificial Intelligence and Machine Learning. It’s a massive, exciting landscape, and I joined DEV because I've heard amazing things about how supportive this community is.

I'm mainly here to learn, connect with other developers, and soak up as much knowledge as I can. If anyone has favorite resources, roadmaps, or advice for navigating the AI/ML space, I’d love to hear them!

Excited to be here! 👋

𝗝𝗼𝗵𝗻 • Oct 15

Welcome to DEV, Emmanuel! 👋

The AI/ML landscape is massive, so the best way to avoid tutorial paralysis is to pick small, hands-on projects early and let practical problems dictate what theory you dig into next.

If you're looking for solid starting points:

Hands-on foundation: fast.ai (Practical Deep Learning for Coders) is fantastic for building working models from day one without getting bogged down in heavy math upfront.
Modern LLM/GenAI workflows: Hugging Face's official NLP course and open-source cookbooks are great for learning real-world pipelines.

Are you leaning more toward the foundational model training/data science side, or building applications on top of LLMs and embeddings?

Emmanuel Gyamfi • Oct 16

Thanks for the warm welcome and the great recommendations, John!
Really appreciate the guidance!

Bohdan Khomyn • Oct 15

Hi everyone! I'm Bohdan from Lviv, Ukraine. I run OLBO, a small web studio, and Balabony, an accessible story platform for Ukrainian readers. I publish about website security and accessibility for small businesses — today: what we learned checking 71 clinic websites for email spoofing protection. Glad to be here!

𝗝𝗼𝗵𝗻 • Oct 15

Welcome to DEV, Bohdan! 👋

Combining web accessibility with practical security audits is such impactful work, especially checking email spoofing protections for clinics, where patient trust is critical.

Across those 71 websites, what was the most common gap you ran into - missing DMARC records entirely, or just misconfigured SPF alignments?

Bohdan Khomyn • Oct 15

Thanks, John!

Missing DMARC entirely, by a wide margin. 56 of the 71 clinic domains (79%) had no DMARC record at all. The other 15 had p=none, so they only monitor. Not one was set to quarantine or reject.

SPF was in better shape: 56 had a record, 15 (21%) had none. To be honest, we didn't test SPF/DKIM alignment, only whether the records exist and what the DMARC policy is, so I can't say how many are misaligned.

We've since run the same check on 215 independent dental practices in London: 10% block spoofed email, 47% have no DMARC at all, 42% are on p=none. Better than Lviv, but the same pattern: the record is either missing or stuck on "monitor".

𝗝𝗼𝗵𝗻 • Oct 15

Those stats are wild, especially the comparison with London.

The "stuck on p=none" trap is so real. Most small organizations set up monitoring, get overwhelmed by raw XML reports they don't have tools to parse, panic about accidentally dropping legitimate appointment or billing emails, and never make the leap to p=quarantine or p=reject.

Having real empirical data across nearly 300 clinical domains is absolute gold. If you haven't already, you should definitely turn these findings into a standalone post on DEV! Data-backed security breakdowns like this get huge engagement here!

CodShot • Oct 15

Hey everyone! 👋

We're new to the DEV Community!

We're the team behind CodShot, a workspace designed to help developers save, organize, and reuse their code snippets.

We're here to connect with other developers, learn from the community, and get honest feedback on what we're building.

Happy to be here! 🚀

𝗝𝗼𝗵𝗻 • Oct 15

Welcome to DEV, team CodShot! 👋

Every developer has a messy graveyard of useful snippets scattered across gists, bookmarks, and random markdown files, so a streamlined workspace to actually organize and reuse them is always welcome.

Are you focusing primarily on a web workspace, or are you planning IDE integrations (like a VS Code extension or CLI) to keep snippets right inside the coding flow?