Moein Kareshk

Machine Learning.
Now @Bloomberg.
Previously Blue Boat Data, @oracle, @dwavesys, @altaml_com, and @UAlbertaCS.

[Moein Kareshk @mkareshk]

@AnthropicAI #AI Releases Bloom: An #OpenSource #Agentic Framework for Automated Behavioral Evaluations of Frontier AI Models

https://t.co/qrIib3zyvh

[Moein Kareshk @mkareshk]

Do you know other well-documented efforts in this area?

[Moein Kareshk @mkareshk]

OpenAI has released a new version of its Preparedness Framework. The framework is indeed important because it helps prioritize and mitigate the most significant risks that accompany advanced #AI capabilities.

Blog post: https://t.co/viZhs2afV5

Document: https://t.co/o9RvJ0FxoC

[Moein Kareshk @mkareshk]

What they're trying to do with this framework is to ensure safety measures keep pace with the rapid evolution of AI.

[Moein Kareshk @mkareshk]

Practical Deep Learning

A free course designed for people with some coding experience, who want to learn how to apply deep learning and machine learning to practical problems.

https://t.co/xA5htngYBo

[Moein Kareshk @mkareshk]

📚 Preparing for machine learning interviews? This guide has topic-focused questions to help you learn & practice efficiently. Created by an LLM-based agent, it's your path to interview success and career confidence! 🚀 #MachineLearning #AI

https://t.co/DfXKvM6dBV

[Moein Kareshk @mkareshk]

Paper: https://t.co/hbQsgt9hGd

[Moein Kareshk @mkareshk]

🛠️📊 Portfawn: A #Python library for #portfolio #optimization using both classical and #quantum techniques. It's designed for flexibility, allowing users to work with customizable objectives, risk models, and optimization methods.
https://t.co/lLOIkKJ5Md

[Moein Kareshk @mkareshk]

Portfawn provides tools for analyzing and visualizing portfolio performance:
- 📈 Cumulative returns.
- 📊 Distribution of returns.
- 🔗 Correlation and covariance matrices.

[Moein Kareshk @mkareshk]

Simultaneous Data Sampling And Feature Selection Via Weak Learners

#MachineLearning #ArtificialIntelligence #FeatureSelection #DataSampling #Innovation #AIResearch #Patent

https://t.co/fWjdSxG2Td

[Moein Kareshk @mkareshk]

Thrilled to share that I’ve joined Blue Boat Data as a Senior Machine Learning Engineer (#GenAI)!
In my new role, I’ll be driving generative AI #research, #development, and #integration. Excited to chart new waters with this amazing team!

[Moein Kareshk @mkareshk]

Today marks the end of my incredible journey at @oraclelabs! I'm deeply grateful for the opportunities, learning, and growth I experienced alongside such talented colleagues.
Looking forward to exploring new challenges and making an impact in my next role!

[Moein Kareshk @mkareshk]

Tutorial on Diffusion Models for Imaging and Vision

https://t.co/tKV34sdnp1

[Moein Kareshk @mkareshk]

🔑 Main Points:
- Similar to Softmax Attention, Sigmoid Attention can also learn any function.
- Sigmoid Attention is more stable during training (handles large gradients better)
- The FLASHSIGMOID implementation makes it faster!
#AI #ML #Transformers

[Moein Kareshk @mkareshk]

Sigmoid Attention! @Apple's recent paper, "Theory, Analysis, and Best Practices for Sigmoid Self-Attention," introduces an interesting alternative to softmax attention in transformer models: Sigmoid Attention.

#Paper: https://t.co/kQeHaElajB
#Code: https://t.co/0ymEmFSAiL

[Moein Kareshk @mkareshk]

Complete the course and assessment to earn the APEX +AI Champion badge. Level up your low-code development skills with AI!
#OracleAPEX #AIinTech @Oracle_Edu

[Moein Kareshk @mkareshk]

🚨 Great Course Alert! 🚨

@OracleAPEX: Empowering #LowCode #Apps with #AI 🚀
🔹 Configure Generative AI Service
🔹 Create AI-powered apps and SQL queries
🔹 Implement AI-driven chatbots
🔹 Utilize AI Vector Search in @OracleDatabase 23ai for #RAG

https://t.co/pAJZQyqECl

[Matthew Guzdial @MatthewGuz]

University of Alberta has a general call for professors in the AI + X space, where X can range from education, the arts, sciences, etc. If you're doing cool work in applications of AI to any area please consider applying!

https://t.co/8heNDkb4X5