Hi, I am Nadav. My primary focus: AI, distributed systems, computer vision and NLP. Currently working @Meta. I co-founded Bzigo, the first consumer device for locating indoor mosquitoes using AI and computer vision. I founded a few companies that help people communicate better using deep networks: eAngel is a human, live grammar correction and proofreading service, cost-effective and with a fast turnaround time. Wordeep is an AI proofreading and writing assistant service, with many integrations: Overleaf, Thunderbird, browser extensions, Google Docs/Slides, Word, and more. At Sun Microsystems (➛ Oracle) we created the Java mobile edition, VM and SDKs. I co-founded I Am Real, a technology for authenticating online identities using social network analysis, computer vision and voice analysis. At mSystems (➛ Sandisk ), we innovated real time cryptographic (elliptic curves, RSA, hashes, etc.) and PRNG algorithms on FPGAs to be used in embedded devices’ controllers. I was the 7th employee in Shopping.com which was later acquired by ➛ eBay. I co-founded Citip, the first pro bono social engine for getting feedback on landlords and the neighborhood, so one can make a better informed decision before renting a property.
I studied BSc. in Computer Science, BSc. in Physics and MBA, all in TLV university. You can reach out to me on Linkedin. My Erdős number is 4.
Back in the nineties I had a highly visited BBS. I wrote the network communication BBS software myself in Pascal + Assembler. I enjoyed experimenting with Blue Boxes. My first programming language, at the age of five, was Logo, and quickly afterwards Basic, then Pascal, Assembler, C, and Delphi. Java, Javascript, PHP, Ruby and Python came only years later.
Patents:
System and method for locating and eliminating insects
Method and system for preventing uploading of faked photos
Posts
- Memory Footprint of a Neural Net During Backpropagation
- PRILoRA: Pruned and Rank-Increasing Low-Rank Adaptation
- Shares Efficient Frontier
- How much will we lose if we buy and sell shares too frequently? Optimal strategy for cashing out shares, tax harvesting and more.
- Distance between two lines in 3D
- Intuitive explanation for the max-min inequality: Why min-max is always greater than max-min.
- ChatGPT - How does it work?
- S4 and State Space Models
- Make a Video - text/image/video -> video
- Simple Gaussian Sampling Algorithm
- Sherlock Holmes: Improving server loading time
- Find the optimal policy in MDP using TD(0) learning and UCB
- Feature-Preprocessing Leakage During Data-Preparation
- Sampling
- Adjacency and Laplacian Matrices of a Graph and Their Special Properties
- XGBoost paper explained
- Square Error and Cross Entropy Loss Derivation using Maximum Likelihood
- Web Product Management and Product Metrics
- Product Development Lifecycle
- Product management | Pricing Methods
- How ESOP Allocating Order Affects the Outcome
- Product Management | Scrum
- Exploding and Vanishing Gradients: On the difficulty of training Recurrent Neural Networks
- Smart Cache: Improvement over memcached
- Immerse Spherical and Hyperbolic Geometry in Euclidean Space Using Gradient Descent