Bio-PrecisionAI Health

11/28/2025

11/28/2025

These 16 GitHub repos are better than any $1,000 AI/ML course:

1:// Machine Learning for beginners by Microsoft (78.5k ⭐)
https://lnkd.in/gggwDP9h

2:// 100 days of ML coding (48.6k ⭐)
https://lnkd.in/gyBQF3dv

3:// All algorithms implemented in Python (212k ⭐)
https://lnkd.in/gdeUgjsi

4:// Mathematics for Machine Learning (14.7k ⭐)
https://lnkd.in/g_TYTu5J

5:// Made with ML (44.1k ⭐)
https://lnkd.in/gMvyzFgK

6:// 60+ implementations of Deep Learning papers (64.1k ⭐)
https://lnkd.in/gR4aC2GQ

7:// Neural Networks: Zero to Hero (18.3k ⭐)
https://lnkd.in/gnusqKFa

8:// Hands-On LLMs book (17.3k ⭐)
https://lnkd.in/gT3diSRV

9:// Prompt Engineering guide (65.9k ⭐)
https://lnkd.in/gmYzhDY7

10:// AI Agents for Beginners by Microsoft (43.8k ⭐)
https://lnkd.in/ghGHGiMk

11:// Generative AI Agent techniques (17.5k ⭐)
https://lnkd.in/gq-c7URx

12:// RAG techniques (22.7k ⭐)
https://lnkd.in/g5j3ksRA

13:// Data Science to learn and apply for real world problems (27.7k ⭐)
https://lnkd.in/gFXr4msv

14:// Awesome Natural Language Processing (17.9k ⭐)
https://lnkd.in/gW9jBJcM

15:// Awesome Reinforcement Learning (9.4k ⭐)
https://lnkd.in/gyXtXQhc

16:// All Reinforcement Learning algorithms from scratch (1.2k ⭐)
https://lnkd.in/g8SdWKJU

Copied

This link will take you to a page that’s not on LinkedIn

11/28/2025

If you want to grow in AI/ML, these are the top 16 blogs worth reading in 2025 👇

These authors train frontier models, deploy AI at scale, and write about what actually works in production:

1. Andrej Karpathy
Deep dives into neural networks, LLMs, and AI fundamentals.
🔗 https://lnkd.in/g6ZgzM7R

2. Chip Huyen
MLOps and production ML systems from someone who's built them at scale.
🔗 https://lnkd.in/gy2HHDgD

3. Sebastian Raschka, PhD
In-depth tutorials on deep learning, LLMs, and ML fundamentals with code.
🔗 https://lnkd.in/gpxCAbnm

4. Interconnects by Nathan Lambert
Analysis of AI policy, RLHF, and the business side of foundation models.
🔗 https://lnkd.in/g_vAQQTb

5. Machine Learning Mastery by Jason Brownlee
Practical tutorials on ML, deep learning, and practical implementation guides.
🔗 https://lnkd.in/gvb5ZR4E

6. Lil’Log by Lilian Weng (ex VP of Research at OpenAI)
Technical breakdowns of RL, transformers, and diffusion models.
🔗 https://lnkd.in/g7Pgwsi9

7. Eugene Yan (Principal Applied Scientist at Amazon)
Applied ML, recommender systems, and data science patterns used in production.
🔗 https://lnkd.in/gNz_wn_B

8. Philipp Schmid (Senior AI Relation Engineer at Google DeepMind)
Guides for deploying LLMs and building AI apps on cloud infrastructure.
🔗 https://lnkd.in/grVwHUMk

9. Hamel Husain (ex GitHub Staff ML Engineer)
MLOps best practices, fine-tuning workflows, and lessons from shipping ML products.
🔗 https://lnkd.in/g34R2tKC

10. Jason Liu
Structured outputs, AI agents, and production-ready LLM patterns.
🔗 https://lnkd.in/gZRYBGYp

11. Berkeley Artificial Intelligence Research Blog
Academic research on RL, robotics, computer vision, and NLP.
🔗 https://lnkd.in/gtaqkPTK

12. Hugging Face
Product updates, tutorials, and the latest from open-source AI.
🔗 https://lnkd.in/g5fJTkRT

13. Google DeepMind
Google's premier AI research division.
🔗 https://lnkd.in/gfwDCfrF

14. OpenAI Research
Latest papers on GPT models, RLHF, safety, and frontier AI systems.
🔗 https://lnkd.in/gG3bWnqY

15. Anthropic Research
Constitutional AI, interpretability, and safety research behind Claude.
🔗 https://lnkd.in/gQ8TATwi

16. Cohere Research
Enterprise LLMs, RAG systems, and production AI deployment.
🔗 https://lnkd.in/ge6wBHdq

Copied

This link will take you to a page that’s not on LinkedIn

11/27/2025

Top 10 Generative AI concepts every AI/ML engineer should know in 2025 👇

(1) Tokenization: https://lnkd.in/gcRG4K_c
Break text into tokens (words, subwords, or bytes) for model input. Tokenization affects context length, multilingual support, and efficiency.

(2) Embeddings: https://lnkd.in/g4gqjKNH
Transforming raw data into dense vectors underpins search, clustering, semantic understanding, and RAG. Embedding space behavior is fundamental to how models “understand” inputs.

(3) Multimodal LLMs: https://lnkd.in/gX4EKtgN
Map different data types (text, images, audio, video) into a shared representation space. Allow models to reason across modalities and align concepts visually and linguistically.

(4) Retrieval-Augmented Generation (RAG): https://lnkd.in/g5tWVhwz
A hybrid approach that combines external search with LLM reasoning. RAG improves factual accuracy, reduces hallucinations, and is a popular strategy in enterprise LLM applications.

(5) Efficient Fine-Tuning (LoRA, QLoRA, PEFT): https://lnkd.in/gyug2CPF
These methods allow parameter-efficient adaptation of large models, reducing compute costs and speeding up deployment. Especially relevant for domain-specific applications.

(6) Model Context Protocol (MCP): https://lnkd.in/gXbMNuz2
Connect models to external tools, data sources, and environments through a standardized protocol. Let LLMs interact with APIs, databases, and applications safely and reliably.

(7) Model Distillation: https://lnkd.in/gWCgqBYK
A technique to compress large models into smaller ones with similar performance. Useful for edge deployment or latency-sensitive applications.

(8) Model Merging: https://lnkd.in/grxybkpm
Combining weights from multiple fine-tuned models can lead to better performance or generalization. An emerging practice for customizing foundation models.

(9) Hallucination Mitigation: https://lnkd.in/gbVKknSY
Reduce fabricated or incorrect answers by adding retrieval, constraints, or validation. Essential for production systems that require factual or domain-safe outputs.

(10) Latent Space: https://lnkd.in/gsU72Y87
A compressed, abstract representation of data where patterns emerge. Navigating this space enables generation, interpolation and creativity in AI.

Copied

This link will take you to a page that’s not on LinkedIn

11/25/2025

AI is about to change forever: Google may have just revealed the biggest shift in AI since the Transformer, and almost nobody is paying attention 😳

Google researchers just dropped something called Nested Learning, and if it works the way early results suggest, it might be the biggest shift in AI since the famous “Attention Is All You Need.”

- Not another bigger model.
- Not another trillion-parameter arms race.
- A completely new learning paradigm.

The core idea is beautifully simple yet mind-bending:

What if AI didn’t learn once but learned continuously? 🤔

- Not just updating a few temporary attention states.
- Not forgetting everything outside a 1M-token window.
- Not freezing after pre-training like it’s stuck in amnesia.

Nested Learning reframes an AI model as multiple learning systems stacked inside each other, each operating at its own speed - some adapting instantly, some slowly consolidating long-term memory.

Exactly like the human brain does.
Fast for experience.
Slow for wisdom.

And Google didn’t just theorize it. They built it 🤖

The proof-of-concept model - HOPE - already beats Transformers and modern RNNs across language modeling, reasoning, and long-context tasks at only 1.3B parameters.

- It retrieves information from massive sequences with ease.
- It resists catastrophic forgetting.
- It evolves during use.

This is not a bigger hammer. It’s a different physics.

→ Instead of stacking more layers, it stacks optimization loops.
→ Instead of making context windows larger, it makes memory smarter.
→ Instead of freezing intelligence, it lets intelligence grow.

If this trajectory holds, Nested Learning could push us closer to something that actually resembles AGI:

↳ Systems that don’t just perform but improve.
↳ Systems that don’t just store knowledge but develop it.
↳ Systems that don’t just follow rules but learn new ways to learn.

And that’s the part people aren’t ready for.

Because once AI models can rewrite their own learning processes, the frontier stops being about scale and starts being about self-improvement.

And that changes everything.

Copied

11/20/2025

Want to deepen your understanding of modern binder design? Join Neurosnap for our annual Binder Design & Analysis Workshop.

In this session, you will learn how to design, evaluate, and optimize protein binders using the latest and most accurate methods in computational biology, including BoltzGen, NeuroBind, BindCraft, RFdiffusion, and molecular dynamics approaches.

We will cover:
- Basics of protein binders and their design
- Different binder modalities and their tradeoffs
- Modern binder design algorithms and when to use them
- Key developability properties to screen for
- Practical strategies to filter, refine, and prioritize candidates

Date: Wednesday, December 10th, 2025
Time: 12:00 PM Eastern Time (US and Canada)
Format: Online webinar

Register here:
https://lnkd.in/e5Dcs-VP

Feel free to share this event with colleagues who work in protein engineering, computational biology, or drug discovery. Spaces are limited so try to register in advance.

Learn how to design, evaluate, and optimize protein binders using the latest and most accurate methods in computational biology — including BindCraft, RFdiffusion, BoltzGen, and molecular dynamics simulations. This hands-on workshop guides you through each step, from sequence generation and struct...

11/11/2025

Well-done, Luper Tsenum!

🌟 Meet the talented team leading Nucleate Florida!
We’re proud to spotlight a dynamic group of scientists, innovators, and changemakers driving collaboration and entrepreneurship across Florida’s life sciences community.
Co-Managing Directors:
• Avni Bhalgat – PhD candidate, Cancer Biology, U Miami | Shapes Nucleate Florida’s vision & strategy
• Olivia Bosquet – PhD candidate, Biochemistry & Molecular Biology, U Miami | Co-leads organizational strategy & partnerships
Communications Team:
• Isabella Altilio – Director of Communications, Nucleate FL | EVP of Content Strategy, Nucleate HQ | PhD candidate, Molecular, Cell & Developmental Biology, U Miami | Researcher at Diabetes Research Institute
• Arnab Chakraborty – Communications & Content Strategy Lead | PhD student, Biochemistry & Molecular Biology, U Miami | Researcher at Interdisciplinary Stem Cell Institute (Cardiology)
Research & Partnerships:
• Leonor Teles – Director of Research | PhD candidate, Biomedical Engineering, U Miami | Bridges science & entrepreneurship
• Louis Cafaro – Director of Partnerships | PhD candidate, Cancer Biology, UF | Builds statewide collaborations
• Elena Bisotto & Mayra Tabares-Beltran – Partnerships & Operations| Ensuring Nucleate Florida thrives internally and externally
New Members:
• Joseph Luper Tsenum – Partnerships Operations Lead | PhD, Biomedical Engineering (AI Drug Discovery), UF | Founder & CEO, Bio-PrecisionAI Health | Founder & Exec. Director, NELIREF
• Sukhmandeep Kaur – Director of Relationship Management | PhD, Pharmaceutical Sciences, FAMU | Builds connections across biotech, consulting, and industry initiatives
• Tyler Sarovich – Partnership Lead | PhD candidate, Neuroscience, FAU | President, Neuroscience Student Organization | Connects cross-disciplinary networks
• Diya Jayram – Activator | MD/MBA student, U Miami | Driving engagement and program activation
• Mainak Bardhan – Partnerships | PhD student, Genomics Department, U Miami | Expanding research and industry collaborations
• Darcy Tocci – Activator | DCI Fellow, Harvard Medical School | Supporting initiatives and community engagement
✨ Together, this team embodies Nucleate Florida’s mission: fostering collaboration, mentorship, and innovation to empower the next generation of biotech leaders across the Sunshine State.

11/10/2025

🚀 We are thrilled to welcome Ellie Okwei, PhD as the Founding Head of Computational Drug Discovery at Bio-PrecisionAI Health!

Ellie completed her PhD in Computational Chemistry and Biophysics, where her research focused on molecular simulations, structure-based drug design, and predictive modeling of protein–ligand interactions. Her expertise spans quantum chemistry, MD simulations, and AI-driven molecular generation—bringing a powerful scientific foundation to our mission.

As Founding Head of Computational Drug Discovery, Ellie will lead our efforts in AI-powered small molecule and biologics design, multi-scale modeling, and integrated chemical–biological data pipelines. She will work closely with our scientific and technical teams to drive innovation at the frontier of generative chemistry and precision therapeutics.

We are thrilled to have Ellie join us as a founding scientific leader and help shape the future of AI-driven drug discovery. Please join us in welcoming her to the Bio-PrecisionAI Health team! 🎉

11/10/2025

🚀 We’re thrilled to welcome Dylan Tan as the Founding CTO of Bio-PrecisionAI Health!

Dylan is currently completing his PhD in Biomedical Engineering at the University of Florida, where his work spans AI/ML, computational biology, and next-generation biomolecular modeling. His deep expertise in multimodal learning, predictive modeling, and large-scale computational infrastructure uniquely positions him to lead our technical vision at the intersection of AI, drug discovery, and precision medicine.

As Founding CTO, Dylan will drive the development of our AI platforms for molecular design, integrate multiomics data pipelines, and help build the technical foundation of our company from the ground up. His leadership and innovative approach will accelerate our mission to create breakthrough therapeutics using advanced generative and predictive AI.

Please join us in congratulating Dylan and welcoming him to the Bio-PrecisionAI Health team! 🎉