The Kickstart
Who am I? Why I've started this newsletter? Whats in for you? What's next ?
Hey,
Thank you for joining Neural Bits, I’m happy to have you on board!
Who am I?
My name is Alex Razvant, I’m a Senior ML Engineer specializing in Deep Learning Systems.
Throughout my career, for over 7 years, I've integrated ML in a variety of use cases, from AI for autonomous vehicles and generative AI (including LLMs and VLMs) to smart surveillance systems, 3D avatar generation, and virtual try-on technologies.
Currently, I'm focusing on applying Vision AI in retail, scaling existing systems, and tackling issues like shrink, checkout, theft prevention, and autonomous shopping experiences.
I had the chance to work with experts in the field, interact with some of the best minds at NVIDIA directly, and see the challenges that Deep Learning Systems impose first-hand - I’ve amassed valuable insights in this area.
Why I Started This Newsletter
While looking for solutions to the challenges I’ve faced I’ve identified gaps, that currently lack expert coverage openly and filling these gap became the main factor of starting the Neural Bits Newsletter.
Working with more classical ML applications like recommender systems or structured data like DataFrames and text files is way easier than handling large volumes of video/image/3D data that Deep Learning implies.
Many solutions and techniques that solve the complex challenges of Deep Learning are either reserved for big players, siloed and not shared, or hidden in theoretical concepts within research papers or certifications that still touch the surface level only - and these don’t necessarily help ML Engineers when they are in need.
Resources are often scarce on practical, applicable insights for real-world challenges when talking about Deep Learning systems.
Of course, you might find how to fine-tune, integrate, and deploy models, or build a product around AI but rarely this is done most efficiently, planning-wise, robustness-wise, and resource-wise
My goal with Neural Bits is to fill these gaps, one bit of practical advice at a time.
Coming with a background rich in working on and deploying real-time Computer Vision Systems - I want to share the lessons I’ve learned.
What’s in for you?
As a subscriber, you’ll get access to expert insights on how to solve challenges that deep-learning projects imply when deployed to production, especially at the Edge.
Here’s a short list of topics on what to expect:
Lessons I’ve learned
Challenges I’ve overcome
Deep dives into Deep Learning Systems
Deep Learning Engineer Roadmaps
End-to-End hands-on courses
MLOps for Vision AI Projects
Deep dives into model optimization (GPUs, TPUs)
Stable Diffusion and Vision Foundation Models
Robotics
Many more
I’m committed to providing you with high-quality, valuable content alongside detailed diagrams, concepts, and code tutorials that’ll help you build and deploy Deep Learning models from 0 to 100.
Regardless of whether you’re a seasoned Engineer, or a beginner trying to get into the field of Deep Learning, I aim for the middle ground:
Unpacking complex concepts and explain them intuitively, such that you could use the insights from my articles - right away.
What’s next?
For the last few days, I’ve been working on setting up this newsletter in a way I could feel satisfied.
Next, I have a multitude of interesting and helpful topics I want to talk about, diving deep into concepts like:
Deploying LLMs at the Edge
Video analytics using Visual RAGs
Multi-modal Models and how to use them
Building a production-ready Computer Vision System
Advanced monitoring of Deep Learning Deployments
Building an AI video classification system to identify Theft Actions
Scaling real-time Vision AI to multiple live cameras
Deep dives into NVIDIA’s Ecosystem
Many more to come
My goal is to share my expertise in building, deploying, and managing the ML lifecycle, while helping you build a strong foundation to understand deep learning systems, ease the learning curve, and get you confident in solving problems and building amazing projects by yourself.
With that being said, I hope it sparked your interest and you’ll join me as I’m excited for what’s to come.
That’s the non-technical bit for today!
Enjoyed this bit? Tell your friends about it :)
Wanna chat? Feel free to message me on 👇 or LinkedIn