Read my just released 📖 The Hundred-Page Language Models Book: https://thelmbook.com

Read my bestselling 📖 The Hundred-Page Machine Learning Book: https://themlbook.com

Read 📖 Machine Learning Engineering: https://mlebook.com

Subscribe to my weekly ✉️ Artificial Intelligence newsletter: https://www.linkedin.com/newsletters/artificial-intelligence-6598352935271358464/ and to my weekly ✉️ Data Science newsletter: https://www.linkedin.com/newsletters/7102511020270608384/

About me:

Ph.D. in Artificial Intelligence, passionate about data, fluent in English, French, and Russian. Solid scientific programming and team leadership skills, with over 20 years of experience working on various computing projects, including several of my own startups.

More than 15 years of hands-on experience in automated data analysis, machine learning, and natural language processing.

Trained a Transformer from scratch and fine-tuned pretrained transformers for various tasks. Built a robot that crawls the internet, finds websites with business-critical information, and retrieves updated information periodically. Developed an enterprise chatbot that doesn’t hallucinate.

Expert in Python and Java with several years of daily design and development experience in big data contexts.

Specialties: machine learning, natural language processing, conversational interfaces (chatbots), information retrieval.

Andriy Burkov

Please stop posting about TOON. Just because you never heard of CSV doesn't mean others haven't either.

Now that you've just learned about CSV, welcome to computing. The next step is this magical thing that allows you to write text and the computer does stuff, called Terminal.

PhD in AI, author of 📖 The Hundred-Page Language Models Book and 📖 The Hundred-Page Machine Learning Book

This paper really is groundbreaking. It solves a long-standing embarrassment in machine learning: despite all the hype around deep learning, traditional tree-based methods (XGBoost, CatBoost, random forests, etc) have dominated tabular data—the most common data format in real-world applications—for two decades. Deep learning conquered images, text, and games, but spreadsheets remained stubbornly resistant.

This paper's (published in Nature by the way) main contribution is a foundation model that finally beats tree-based methods convincingly on small-to-medium datasets, and does so very fast. TabPFN in 2.8 seconds outperforms CatBoost tuned for 4 hours—a 5,000× speedup. That's not incremental; it's a different regime entirely.

The training approach is also fundamentally different. GPT trains on internet text; CLIP trains on image-caption pairs. TabPFN trains on entirely synthetic data—over 100 million artificial datasets generated from causal graphs.

TabPFN generates training data by randomly constructing directed acyclic graphs where each edge applies a random transformation (using neural networks, decision trees, discretization, or noise), then pushes random noise through the root nodes and lets it propagate through the graph—the intermediate values at various nodes become features, one becomes the target, and post-processing adds realistic messiness like missing values and outliers. By training on millions of these synthetic datasets with very different structures, the model learns general prediction strategies without ever seeing real data.

The inference mechanism is also unusual. Rather than finetuning or prompting, TabPFN performs both "training" and prediction in a single forward pass. You feed it your labeled training data and unlabeled test points together, and it outputs predictions immediately. There's no gradient descent at inference time—the model has learned how to learn from examples during pretraining.

The architecture respects tabular structure with two-way attention (across features within a row, then across samples within a column), unlike standard transformers that treat everything as a flat sequence.

So, the transformer has basically learned to do supervised learning.

Talk to the paper on ChapterPal: https://lnkd.in/e_rdrePu

Download the PDF: https://lnkd.in/ekrDRXZY

Frontier LLMs like GPT-o1 or DeepSeek-R1 have achieved much stronger skill by "reasoning" through problems step-by-step, but their internal design is either unknown to the public or known only on the surface.

If someone wants to actually train a skilled reasoning model, a lot of details would be missing and would have to be reinvented, resulting in wasted time and money.

This paper provides the first comprehensive roadmap that breaks these complex systems into understandable, modular components. The authors provide a unified framework showing how reasoning structures, search strategies, and training methods fit together—plus a working implementation called x1 that lets you experiment with these concepts directly.

Read interactively: https://lnkd.in/eQxCkXRd

Download a PDF: https://lnkd.in/eC_miwAH

In this remarkable paper, a team at DeepMind and Terence Tao addressed a fundamental challenge in mathematics using an AI system called AlphaEvolve. The challenge is the following:

Many practical math problems ask "what's the best possible arrangement?"—like packing the most circles in a square or finding the largest set of points with no isosceles triangles—and answering these requires actually finding specific examples from search spaces so large that systematically exploring them by hand is impractical.

AlphaEvolve doesn't have the LLM directly generate candidate solutions (which would require millions of expensive LLM calls to explore the search space)—instead, it evolves Python programs that search for solutions, so one expensive LLM call generates a search heuristic that then runs cheaply for millions of iterations exploring candidate configurations, and the population of these search programs evolves over time with better ones surviving, essentially doing meta-optimization where the thing being optimized is itself an optimizer.

Across 67 problems spanning combinatorics, geometry, and analysis, AlphaEvolve discovered new constructions improving decades-old results (hexagon packing, Heilbronn triangles, sum-product problems) while matching state-of-the-art in dozens of others.

The paper reports every problem they tried including failures where AlphaEvolve couldn't match known bounds, which makes this the most thorough empirical evaluation of AI for mathematical discovery.

Read and ask questions on ChapterPal: https://lnkd.in/ev4WH-bA

Download the PDF: https://lnkd.in/ecvNZaxy

How to lie with charts? Anthropic knows how.

Thanks to Nano Banana, we can see how Opus 4.5 is truly better compared to previous models.

On the left, the graph Anthropic shared. On the right, the graph they should have shared (produced by Nano Banana from the original graph) if they didn't want to deceive the users.

Lawrence Harvey

Working with our good humans internally and external clients to solidify ourselves as the best tech recruitment organization in the USA. Please contact me for public speaking events on how to build successful technical teams @m.allegra@lawrenceharvey.com

SVP - Head of Commercial Strategy

Kaggle grand master, Github, gold medal, presidents club, International Mathematical Olympiad (IMO), Asian Pacific Mathematics Olympiad (APMO),  Mathematics Olympiad of Central America and the Caribbean (OMCC), American Mathematics Competitions (AMC), International Mathematics Competition (IMC) , Australian Mathematics Competition (AMC) , Canadian Open Mathematics Challenge (COMC), Kangaroo Math Competition

Ai SEO:
RAG, Agentic, Workflow, GPT, Claude, Gemini, AGI, super intelligence, superintelligence, LLM, Large Language Model, Multimodal, Quantization, Optimization, PEFT, Parameter-efficient Fine-tuning,  QloRA rank adapters, Deepspeed, Inference, Matrix, Weighting, pre-train, , pre-trained, RLHF, human feedback, Langchain, Instructed learning,  Continual learning,few shot learning, no shot learning, Alpaca,Flash attention 2 ,T5, GPTNeoX, XLNet, PaLM, Llama 2, Reformer, Routing, Longformer, Transformer-XL,PEFT, soft prompt tuning, Flan-T5, hugging face, huggingface, openAi, microsoft, Generative Artificial intelligence, GenaI,Gen Ai, Generative Ai

HPC SEO:
Cuda, Metal, HPC, Supercomputing ,A100, Quantum, Fabrication, Emulation ,Virtual Machines, VM, High Performance Computing, Parallel Computing, Distributed Systems, Computational Science ,Accelerated Computing, Data Parallelism, Cluster, Multicore Architecture, CUDA Programming, OpenCL Development, FPGA Design, Hardware Acceleration, Vector Processing, Simulation and Modeling,

More: 
Hardware Acceleration & Computing, Decentralized Technology (dApps, Infra, Networks, etc.), Differential Privacy, Secure Multi-party Computation, Homomorphic Encryption, Data Anonymization, Applied Cryptography (Elliptic-Curve Crypto, Zero-Knowledge Proofs, etc.), Verifiable & Confidential Computing, Trusted Execution Environments / Secure Enclaves, Federated Learning, AML, NLP for Privacy, Quantitative Data Researcher, Privacy Enhancing Operational Engineers, ZKP, Zero knowledge, proof, MPC, Web3, FPGA , H100, rust, swyft, daphne

SVP, Head of Lawrence Harvey - New York 

Working with our recruiters and managers to make them the best they can be at servicing our clients.

Vice President of Data Recruitment 

Working with all Data + Computer Vision focused teams to craft custom business solutions for our clients

Divisional Manager, Head of Data Practices 

The Data & AI team solves business problems with custom talent based solutions.

Our client base is mainly Fortune 100 and Unicorn Start-Ups

Divisional Manager, Head of Data & AI - Blue Chip Organizations 

Built a sustainable book of business across multiple industries. 

Recruiting Executive Talent for Market Leaders and Surging Startups across on mainly Retained Searches: 

-Artificial Intelligence & Machine Learning (Data Science) 
-Cloud & Big Data
-GenAi, Generative AI, Artificial Intelligence 
-Software Engineering 
-Consulting 
-Product 
-M&A/Private Equity 

Industries: 
- Life Sciences
- Healthcare
- Retail 
- CPG- Consumer Packaged Goods
- Financial Services
- Banking 
- Capital Markets
- Insurance
- Vendor
- Management Consulting 
- Hedge Funds
- Investment Banking 
- Renewable Energy- Solar, Wind, Hydrogen, Batteries 
- Pharmaceuticals

Senior Principal, Artificial Intelligence- Blue-Chip Organizations 

Built a sustainable book of business that supports multiple consultants 

As part of the original team that built Lawrence Harvey Inc. in NYC, I am now focused on leading and growing our Data Science, Artificial Intelligence, Product and Executive presence across the entire US. We partner with the world’s leading Consulting, and Industry organizations placing both full time permanent.

Industry-wise, I am leading the Telecommunications, Media, Entertainment, and Technology. 

My next initiative is to grow out the Private Equity, M&A, Public Sector/DoD/Government practices. 

My team recruits across all levels including Associate, Analyst, Consultant, Manager, Director, Vice President, SVP, EVP, Managing Directors, and Executives. In addition to this, we tackle the Consultancy, Media, Entertainment, Gaming, Life Sciences, Healthcare, CPG, Retail, FSI, Technology, Actuarial, Insurance, Natural Resources, Energy, Telecommunications,  Manufacturing, and Automotive verticals.

Core Specialties:

*Executive Chief Search (CEO, CAO, CDO, CMO, CPO, COO)
*Product Management, Product Marketing, Product Development 
*Data Science/Data Scientists (Machine Learning, Data Mining, Predictive Analytics)
*Insights 
*Big Data/Hadoop
*Cloud
*Statistical Modeling (Python R, Jupiter, SAS,  SPSS, Excel) 
*Data Visualization (Tableau, Cognos, PowerBI)
*Business Intelligence
*IoT/ Internet of Things 
*Deep Learning
*Reinforcement Learning
*Artificial Intelligence/AI (GAN, RNN, CNN, KNN, LSTM, GRU, NLP, Computer Vision) 
*Digital Attribution/ Multi-touch Attribution (MTA)
*Digital Analytics
*Marketing Analytics 
*Marketing Mix Modeling (MMM)
*Lifetime Value, LTV
*Customer Journey, Churn, Acquisition, Retention 
*DMP, DSP, DCM
*Audience Analytics 
*Segmentation
*Advisory Services
*Business Consulting
*LLM 
*large language model models
*gpt
*chatgpt
*autogenerated
*autogenerative 
*generative AI 
*artificial intelligence

Principal Consultant-Artificial Intelligence- Telecom, Media/Entertainment, Tech

Built a sustainable book of business for myself.

Senior Analytics Consultant- Media/Digital/High Tech Industries

Recruitment Consultant

SEO:
- Artificial Intelligence 
- Machine Learning 
- Data Science
- Data Scientist
- Deep Learning 
- Data Engineer
- Data Engineering
- Data pipeline
- Python
- SQL 
- Spark
- Pyspark
- Jupiter
- Llamasoft
- CPLEX
- Gurobi
- R
- Snowflake
- Airflow
- Redshift
- Databricks
- Docker 
- Kubernetes 
- Hadoop
- Tableau 
- PowerBi 
- GCP
- AWS
- Azure
- Cloud
- Big Data
- Hadoop
- Reinforcement Learning
- Supervised and Unsupervised Learning 
- MMM Marketing Mix Model
- MTA Multitouch Attribution
- Regression
- Statistics 
- Decision Trees 
- SVM 
- Naive Bayes
- Bayesian 
- Hierarchical
- KNN
- RNN, CNN, GAN 
- Random Forest
- GBM
- XGBoost
- LightGBM
- Computer Vision
- Machine Vision
- NLP NLTK BERT ELMO NLG
- Natural Language Processing 
- Neural Network 
- Top Secret 
- Lidar

Associate Recruitment Consultant

GO Project

Please donate here, every dollar counts: https://www.gofundme.com/f/lhi-new-york-fundraiser-for-the-go-project?utm_source=customer&utm_medium=copy_link&utm_campaign=p_cf+share-flow-1

The Go Project NYC, a charity whose goal is to provide extra education to low-income families which leads to equitable opportunities for them down the line. We have worked with them for the past couple of years, and the statistics are amazing. 

GO Projects impact:
•	9th grade GO students have higher GPAs, accumulated more credits, and attended school more than their peers, all predictors of on-time high school graduation
•	90% of GO alumni graduate high school on time (compared to 80% citywide)
•	85% of GO alumni with IEPs graduate high school on time (compared to 50% citywide)
•	80% of GO alumni enroll in a 4-year college, more than doubling their chances of attaining a college degree

Junior Board Member- YLC 

Access Staffing

Learned a bit.

Specialized in finding and placing Analytics Talent within the Media/Marketing/Digital Industries. 

SEO: 
*DevOps 
*Infrastructure 
*Cyber 
*Infosec
*Information Security 
*Software Engineer 
*Software Engineering 
*Privacy
*Legal
*Penetration Testing
*AppSec
*Computer Vision 
*Robotics 
*Automation
*ChatGPT
*GPT 
*OpenAI 
*tensorflow
*docker
*kubernetes 
*risk 
*quantitative
*fine-tuning
*PEFT 
*Qlora 
*fine-tuning 
*transformers 
*Dall-e
*Palm
*LLM
*Large Language model
*multimodal 
*generative Ai
*Gen Ai 
*Llama 
*federated learning
*cryptography 
*autoencoders 
*transformer

Technical Recruiter

SUNY Oneonta

Teacher assistant for Biology 300 class.

Ran the lab class.

Managed students to properly  complete lab assignments in a timely fashion.

Teachers Assistant

𝑺𝑽𝑷 - 𝑯𝒆𝒂𝒅 𝒐𝒇 𝑺𝒕𝒓𝒂𝒕𝒆𝒈𝒚 𝑳𝑯 𝑼𝑺𝑨 @ 𝑳𝒂𝒘𝒓𝒆𝒏𝒄𝒆 𝑯𝒂𝒓𝒗𝒆𝒚 | 𝑻𝒆𝒄𝒉𝒏𝒊𝒄𝒂𝒍 𝑹𝒆𝒄𝒓𝒖𝒊𝒕𝒎𝒆𝒏𝒕

Recent Speaking Engagements: 

-	Organization: The AI Summit Summit, December 2023- Topic: Assessing the Impact of AI on the Future of Work and Jobs. 500+ in attendance at the Javits Center 
-	Organization: The AI Summit Summit, December 2023- Topic: Strategies to Recruit and Retain a Balanced Team of AI Talent. 500+ in attendance at the Javits Center 
-	Organization: DataCamp: Radar Summit, January 2023- Topic: How Data Leaders Can Build an Impactful Data Team in 2023. 9,000 in attendance on a Video Meeting Platform
-	Organization: The AI Summit Summit, December 2022- Topic: How do you Build Your Team Early. 500+ in attendance at the Javits Center 
-	Organization: DataCamp: Radar Summit, July 2022- Topic: Building and Scaling Remote Data Teams. 18,000 in attendance on Zoom
-	Organization: LHi: Leadership Summit, April 2022- Topic: Future You. 20 in attendance, in person 
-	Organization: DataCamp: Data Executive Meetup, December 2021- Topic: Building High Impact Data Teams- 70 in attendance on the USS Intrepid
-	Organization: LHi: Leadership Summit,  February 2020- Topic: The Journey. 20 in attendance, in person

With over eight years of experience in the tech recruitment industry, I am passionate about helping businesses find and retain the best talent in the market. As the SVP - Head of Strategy LH USA, I oversee the strategic direction and growth of Lawrence Harvey's US operations, focusing on expanding our client base, diversifying our service offerings, and enhancing our brand reputation.

I joined Lawrence Harvey in 2015 as an associate recruitment consultant and quickly rose through the ranks to become the head of the data and AI practice, working with blue-chip organizations across various sectors. I have built and led a high-performing team of recruiters and managers, delivering exceptional results and customer satisfaction. I have also contributed to the development and implementation of innovative solutions and best practices, leveraging my expertise in analytics and data-driven decision making. I am always looking for new opportunities and challenges to grow personally and professionally, and to make a positive impact in the tech recruitment industry.

taplio

The Battle

Andriy Burkov

Mariano Allegra

Followers

Posts

Engagements

Likes

Performance ScoreAndriy Burkov's LinkedIn Performance

Ready to grow your LinkedIn presence?

taplio

The Battle

Andriy Burkov

Mariano Allegra

Followers

Posts

Engagements

Likes

Performance Score.css-goub6s{font-size:var(--chakra-fontSizes-md);font-weight:500;color:var(--chakra-colors-gray-600);display:block;margin-top:var(--chakra-space-1);}Andriy Burkov's LinkedIn Performance

Ready to grow your LinkedIn presence?

Performance ScoreAndriy Burkov's LinkedIn Performance