Break Into Data

Hold your seats tight! Tomorrow, I'll be launching the first podcast episode of 'Technical Founder,' featuring Jonathan Cornelissen, CEO of DataCamp. In only 60 minutes, we unpacked: - How to stand out in this job market for data and AI roles. - How Datacamp became the largest edtech company in Data and AI. - How AI is shaping the present and the future of the workforce. I gained fresh, unfiltered insights into what it takes to build an educational platform used by 14 million people ?? worldwide. I personally enjoyed learning how Datacamp helps early learners overcome beginner's hell and imposter syndrome with their unique hands-on approach. Sign up on merinova.substack.com to get notified once the episode launches tomorrow! ... I'm grateful to have Jonathan as my first guest at the "Technical Founder" podcast. It's been an honor to learn from his journey. Hopefully, you will enjoy the episode as much as I did.

5 Best Github repos to help you pass ML and AI Interviews: 1. Machine Learning Interviews from MAANG (8.1k stars): https://lnkd.in/gq_huuZD 2. 100 Days of ML code (43k stars): https://lnkd.in/gfWZfaDa 3. Machine learning Cheat Sheet (6.4k stars): https://lnkd.in/gAJ3x4Kh 4. System Design Primer (252k stars): https://lnkd.in/gUEqUpAv 5. Python Algorithm Implementation (178k stars): https://lnkd.in/g2xPx_tu If you find these helpful... ?? React ?? Share ?? Comment So more people can learn.

Top AI and ML experts come from diverse backgrounds. Here are my favorites who I follow religiously. ?? ??Disclaimer: Following these creators may cause severe fomo and an uncontrollable urge to learn. Explore their content at your own risk. Deep Learning:? - Sebastian Raschka, PhD, author of the "Build a Large Language Model from Scratch" book. - Kostya Numan,?CTO at?Break Into Data?and founder of the AGI research lab,?Numan.ai . - Karun Thankachan - Senior Data Scientist at Walmart and RecSys researcher. Machine Learning: - Chip Huyen - author of "Designing Machine Learning Systems"(O'Reilly 2022)?and VP of AI at Voltron Data. - Aurimas Griciūnas, founder of Neptune.ai, focusing on MLOps solutions for machine learning projects. - Santiago Valdarrama, founder at Tideily and ML educator on Youtube. - Krish Naik, educator who provides tutorials on ml and deep learning with real-world scenarios. ML and AI engineering: - Aishwarya Naresh Reganti, tech lead at AWS and educator on LLMs and GenAI. - Andriy Burkov, author of the "100 days of ML" book. - Pau Labarta Bajo, real-world ML systems educator and content creator. AI product management:? - Marily Nika, Ph.D, Gen AI product lead at Google and AI Product Management Educator. - Allie K. Miller, US Head of AI Business Development for Startups at Amazon, known for her contributions to AI product strategy MLops or LLMops:? - Maria Vechtomova, MLops tech lead and Databricks beacon. ? - Rapha?l Hoogvliets, Tech Lead ML engineering and content creator. - Paul Iusztin, Founder at DecodingML and educator. Developer Advocacy:? - Megan Lieu, Linkedin content creator on Data Science career. - Sonam G., Podcast host and Community leader. - Aishwarya Srinivasan, Senior AI advisor and Linkedin Top Voice. ... Who else should we add to the list? Share in the comments. ?? #machinelearning

If you want to get a job as an AI engineer, don’t start with the hottest tools and frameworks like Langchain, Llamaindex or Pytorch. Sure, you’ve probably heard everyone talking about the latest AI tools and frameworks like they're the secret sauce to success… but here’s the truth nobody tells you. Technologies come and go—what’s cutting-edge AI or LLM tool today might be outdated tomorrow. But! Fundamentals underneath these tools stay forever. You need to learn to walk before you can run! Focus on these first: ?? Understand Neural Networks LLMs are built on neural networks, so getting a handle on the basics like layers, activation functions, and backpropagation will take you further than focusing on the latest AI fad. Start here: Neural Networks and Deep Learning - https://lnkd.in/gsPi-nSj ?? Prompt Engineering Basics LLMs are only as good as the prompts they receive. Learning the fundamentals of crafting effective prompts can help you get better results from any AI coding assistant or LLM tool. Start here: Prompt Engineering Guide - https://lnkd.in/gtVtw5nU ?? Data Curation & Preparation LLMs need vast amounts of data to perform well. Understanding how to clean, preprocess, and organize data will always be a high-demand skill in this space. Start here: Training & Fine-Tuning: Data -https://lnkd.in/gmMApezn ?? NLP & LLM Fundamentals Natural Language Processing (NLP) is the heart of LLMs. Get familiar with concepts like tokenization, embeddings, and attention mechanisms to understand what drives LLM performance. Start here: Andrej Karpathy’s Zero to Hero playlist - https://lnkd.in/gQuHDJup ?? Model Deployment & Scaling Knowing how to deploy, fine-tune, and monitor LLMs in production is just as important as building them. Cloud infrastructure and APIs will be your best friends here. Start here: Large Language Models: Application through Production - https://lnkd.in/gn8fBrrJ Tools and technologies will continue to change, but these foundational skills and understanding of the fundamental deep learning concepts will stay forever and will help you adapt to any additional layers of GenAI or LLM-related tools. I am hosting a live session with IBM and Clicked to share more on the current AI and LLM landscape, as well as career trajectories, so you can learn how to position yourself well in this rapidly growing field. ?? Date: September 24th. ?? Hurry and sign up, while the registration is still open! https://lnkd.in/gKc5vqS8 #AICareers

When the creator of Pandas, Wes McKinney, was asked to give advice for Data Scientists, he had a shocking response: " Learn SQL ?? I didn't believe what I heard. The creator of one of the largest Python libraries with 60 million downloads every month, tells everyone to learn SQL! I chuckled but was intrigued to hear his explanation. Here is what he said: " Well, some people on the Discord are going to laugh because I'm going to say that learning SQL is actually a really good skill. It's not just learning SQL the language, but learning the concepts of relational algebra and how to think about data sets, designing schemas, and organizing data." ... "It is about learning the file formatting and the basics of data storage, data partitioning, and the relationship between the execution engines. All of these things will yield you to be a better DBT user, a better Snowflake user or a Databricks user. ... "Ultimately, it might seem like the wizard behind the curtain, but I think a lot of these systems have very similar architectures and properties. This will help you understand how these systems work and how to use them effectively." ... I loved the advice and we have a lot to learn from him. Why? Because of his impressive contributions and his background: ?? Wes McKinney is a notable figure in the open-source community! - He dropped out of college to develop the Pandas library. - He co-founded the Apache Arrow project, a cross-language development platform for in-memory data. - He developed the Ibis project, which provides a more efficient way to bridge SQL and Python for data analysis. - He published O'Reily book -?"Python for Data Analysis," a widely-used reference book for learning Pandas. ... If you haven't heard about him or his work yet, I highly encourage you to register for tomorrow’s Live session! Where you can ask questions and learn from his journey. Register here ?? https://lu.ma/i80kh30z

You no longer need to choose between 4 different LLM approaches on your own. I’ve finally published a long-awaited article that breaks down the most common LLM implementations in the industry. The goal is to help guide you with short-term technical decisions and to build your long-term vision for your ML career! You can read it here https://lnkd.in/gZJ9PC3i #machinelearning Would love to hear your thoughts and more topic suggestions!

I should probably NOT say this, but there is a hidden hierarchy in LLM engineering that nobody talks about. The truth is, there are 4 levels of LLM implementations that impact your compensation. Companies are well aware of this and allocate their budgets according to each level. To truly succeed, focus on one and position yourself as an expert. Here are 4 main LLM expertise levels: 1. AI Enthusiasts: Basic understanding of how LLMs work. Passionate about AI, but limited practical engineering experience. Compensation: Entry-level salaries or internships. $60,000 - $90,000 2. AI Engineers Don’t work with underlying architecture of the models but know how to integrate and deploy LLMs with external databases and architectures like RAG through API calls. Compensation: High-end salaries. $100,000 - $200,000+ 3. ML Engineers Deep expertise in machine learning algorithms and optimization. Develop custom LLMs with advanced fine-tuning methods. Compensation: Competitive industry salaries, often with equity. $200,000 - $300,000+ 4. ML Researcher Engineers Contribute to cutting-edge LLM research and publications Develop novel architectures, training techniques, or applications. Bridge the gap between academic breakthroughs and industry implementation. Build models from scratch. Compensation: Top-tier salaries, significant equity, and research grants. $300,000 - $500,000+ ... Tomorrow, I will release an article with more details on each level. So you will understand the LLM industry better and build your portfolio accordingly. Sign up here to get access - merinova.substack.com. ... What level are you excited about the most? #machinelearning #llm P.S. Keep in mind that compensation numbers are a ballpark estimate. The final numbers depend on your location.

15 Big Tech company blogs that will help you become a better ML / AI Engineer: ?? 1. Meta AI - https://lnkd.in/gGMp-Jh2 2. Netflix Recommender System - https://lnkd.in/gm4pTRf4 3. Google AI research applications - https://lnkd.in/gctN7Ths 4. NVIDIA Data Science - https://lnkd.in/ghzhBPnm 5. Apple ML Research - https://lnkd.in/gcJggDju 6. Stripe ML for Fraud Detection - https://lnkd.in/gKS4F-V3 7. Databricks Data Science & ML - https://lnkd.in/gRPE8Gbm 8. Uber AI - https://lnkd.in/gXA8UEBU 9. Grammarly NLP/ML - https://lnkd.in/guMDtPfW 10. Pinterest Ads Recommender - https://lnkd.in/g8QXRH3i 11. Airbnb AI & Machine Learning - https://lnkd.in/gzAXfQg5 12. Microsoft ML - https://lnkd.in/gm-aSSP9 13. DoorDash Data Science and ML - https://lnkd.in/gHkDwpvC 14. MongoDB AI - https://lnkd.in/g8c3HNaa 15. Amazon Machine Learning Blog - https://lnkd.in/g2Q3ZmEh ———— If you find these helpful... ?? React ?? Share ?? Comment So more people can learn.

There are only 5 Machine Learning paradigms that dominate the industry today: 1. prediction: it looks at past data, adjusts for trends, and smooths out random changes to make predictions. (eg, Arima, LSTM) 2. classification: it finds patterns that separate one thing from another. (XGBoost, logistical regression, SVM) 3. generation: it creates new things only from the things it saw before. (GANs, Transformers) 4. clustering: it groups similar things together without knowing what they are. (K-means, hierarchical clustering) 5. recommendation: it suggests things based on your past behavior around similar things you liked. (Two-Tower models, collaborative or content-based filtering) ... Which one do you want to master the most? #machinelearning

I just hit 50k followers. Here is exactly how I did it... (without buying a LinkedIn course) 8 months ago I wrote my first post. But only in the last 3 months, I started to see crazy growth. What has changed? It is a secret that can be applied to everything else in life. I hate to break it to you. But consistency is not enough... Here is what really matters: - Empathy: learn about your reader's hopes & dreams and speak to them. - Mindset: wake up with the burning desire to win and get ready to work for it. - Experiment: fail, learn, and repeat and you will find your voice. Every single time I post, I follow these rules. I don't have content 'templates' or 'frameworks'. But I follow the above religiously. Because of that, I have been able to: - Land 5-figure partnership deals (stay tuned) - Sell out our first ML workshops with Break Into Data - Meet inspiring and energetic people that only add to my life. LinkedIn growth is all about delivering the right message to the right people. It is simpler than you think. We are building a writing tool to help you to do just that. If you want to sign up for the free alpha version release, Follow the link here - https://lnkd.in/gWfWuSfQ ... What is your strategy for growth?