Sattvik Sahai

Applied Scientist at Amazon AGI (Opinions are my own)

[Sattvik Sahai @SahaiSattvik]

Heading to San Diego for #ACL2026! 🌴 I’ll be presenting our paper

📘 From Narrow Unlearning to Emergent Misalignment: Causes, Consequences, and Containment in LLMs. https://t.co/uwuIVcTdOs

Let's connect if you're attending and want to chat about AI safety, and Agent Evaluation

[Amazon Science @AmazonScience]

Calling Trusted AI researchers: Amazon is accepting poster submissions for a one-day symposium on January 21. Papers using #AmazonNova models get priority. Submit by December 31:

[Amazon Science @AmazonScience]

Applications are open for the #AmazonNovaAIChallenge. University teams are invited to advance trusted, useful agentic AI for software engineering:
https://t.co/qdbSiDQRr8

[Amazon Science @AmazonScience]

Announcing a private AI bug bounty program to strengthen the Amazon Nova foundation models. Building on 30+ findings and $55,000+ in rewards from the public program, the new track partners with security researchers and academics to strengthen AI security. https://t.co/OcqEVWxnDY

[Rahul Gupta @rahul1987iit]

🚀 New paper alert: “Amazon Nova AI Challenge — Trusted AI: Advancing secure, AI-assisted software development”

We ran a global university competition—five teams built red-teaming bots, five built safe AI coding assistants—to push forward multi-turn safety alignment.

[Rohit Prasad @RohitPrasadAI]

Last month, we celebrated the winners of our first Amazon Nova AI Challenge and I just have to say, the future of AI is in great hands. These students brought impressive technical skills and fresh thinking to some of AI's toughest problems. Their research papers are worth reading – great insights on responsible AI development: https://t.co/fkMR6MwV8f

[Amazon Science @AmazonScience]

Meet the champions of the Amazon Nova AI Challenge, where university teams around the world went head-to-head to break and defend LLMs under real adversarial pressure:
🏆 Defending: UIUC (PurpCorn-PLAN)
🏆 Attacking: Purdue (PurCL)
🥈 CTU Prague & Nova Lisbon

These students just raised the bar for secure Gen AI.

Full recap: https://t.co/EaP0maph1c

[Amazon Science @AmazonScience]

8 top university teams. One goal: advance secure, AI-assisted coding.

The Nova AI Challenge finals pushed boundaries in red teaming, reasoning, and model safety.

�� Winners announced on July 23
🎥 Watch an inside look at the finals with our judges:

[Amazon Science @AmazonScience]

Eight university teams advance to the Amazon Nova AI Challenge: Trusted AI Track finals, competing head-to-head to strengthen security in AI coding assistants. #ResponsibleAI https://t.co/3AVVHrRZyr

[Amazon Science @AmazonScience]

The Amazon Nova AI Challenge brings together top university teams to accelerate the field of genAI. Ten teams are competing in this year's challenge to advance secure AI-assisted software development, with investments exceeding $5m in stipends, prizes & AWS credits. Learn moreThe Amazon Nova AI Challenge brings together top university teams to accelerate the field of genAI. Ten teams are competing in this year's challenge to advance secure AI-assisted software development, with investments exceeding $5m in stipends, prizes & AWS credits. Learn moreThe Amazon Nova AI Challenge brings together top university teams to accelerate the field of genAI. Ten teams are competing in this year's challenge to advance secure AI-assisted software development, with investments exceeding $5m in stipends, prizes & AWS credits. Learn moreThe Amazon Nova AI Challenge brings together top university teams to accelerate the field of genAI. Ten teams are competing in this year's challenge to advance secure AI-assisted software development, with investments exceeding $5m in stipends, prizes & AWS credits. Learn more:

[Amazon Science @AmazonScience]

Today, Amazon announced its Trusted AI Challenge, which involves university students competing to securely advance large language models (LLMs) that code, and where $700,000 in cash prizes will be allocated across the top four performing teams. Learn more about the challenge and how to apply: https://t.co/aPfPg4OAo4 #ResponsibleAI #GenAI

[Sattvik Sahai @SahaiSattvik]

@anmarasovic Hey Ana, I’d be happy to help if you still need a reviewer.

[Amazon Science @AmazonScience]

The teams selected for the challenge, which began this month, include five returning competitors and four new universities. Learn more about them, and how they hope to be the first to win the $1 million research grant. #AlexaPrize

[Amazon Science @AmazonScience]

Applications for the #AlexaPrize TaskBot 2 are now open! University teams will compete to develop multimodal conversational agents that assist customers in completing complex tasks, with a $500,000 prize for first place. Deadline to apply is November 4th. https://t.co/PfDRR3JA9g

[Sattvik Sahai @SahaiSattvik]

@iamtrask DM'd you

[Sattvik Sahai @SahaiSattvik]

@AndrewYNg @deeplearningai_ @coursera @pgashwin