Product manager for the Automated Reasoning team at @AWScloud. Interested in #motorbike travel and #mountainbiking - opinions are my own
Vancouver, BC
Joined November 2007
- Tweets8.2K
- Following887
- Followers2.8K
- Likes3.3K
A modern remake of classic Baldur's Gate games?! Yes please @Hasbro, @Wizards, take my money now! polygon.com/baldurs-gate-2-r…
Stefano Buliani retweeted
PitCrew built AI agents for financial services that can mathematically prove their outputs are correct.
Every output is verified against regulations and business policies using neurosymbolic AI.
Not probably correct. Provably correct. With a full audit trail.
-30 minutes instead of 2 weeks for manual cross-referencing
-10 minutes to onboard clients, down from 3–4 hours
-30 seconds to clear marketing posts vs. a 3-day compliance review
Your experts go back to doing expert work. Built with Automated Reasoning checks in Amazon Bedrock Guardrails. go.aws/4eICHWO
Stefano Buliani retweeted
PitCrew built AI agents for financial services that can mathematically prove their outputs are correct.
Every output is verified against regulations and business policies using neurosymbolic AI.
Not probably correct. Provably correct. With a full audit trail.
30 minutes instead of 2 weeks for manual cross-referencing
10 minutes to onboard clients, down from 3–4 hours
30 seconds to clear marketing posts vs. a 3-day compliance review
Your experts go back to doing expert work.
Built with Automated Reasoning checks in Amazon Bedrock Guardrails
youtu.be/n9hb3Zg6e34?si=JJTP…
Stefano Buliani retweeted
Amazon Logistics cuts EV charging design reviews from 8 hours to minutes.
Automated Reasoning checks in Amazon Bedrock Guardrails automate reviews and deliver formal verification on every compliance determination with complete transparency and auditability. Engineers can now focus on judgment calls while the system scales efficiently.
go.aws/4ctm0yU
Stefano Buliani retweeted
Reviewing & refining Automated Reasoning policies just got easier. 🔍📋🎯
#AutomatedReasoning policies now include references to the source document. With up to 99% accuracy at detecting correct responses from LLMs, delivering provable assurance in detecting AI hallucinations. #AWS #AmazonBedrock
👉 go.aws/4aJBXAd
Stefano Buliani retweeted
What if your chatbot could explain its reasoning with mathematical proof? 💡🔐📐
Follow this reference to use Automated Reasoning checks to iterate on chatbot responses with precise answers & verifiable proofs. #AWS #AutomatedReasoning #AmazonBedrock
👉 go.aws/4aot0us
ALT Building chatbots with Automated Reasoning checks
Stefano Buliani retweeted
Speed up Automated Reasoning policy validation with conversational refinement ⚡👻🔍
Use Kiro CLI to ask questions about your policies, understand behavior & address issues. #AWS #AmazonBedrock #generativeAI @kirodotdev
👉 go.aws/4aetXan
ALT Use Kiro CLI to refine Automated Reasoning Policies
Get in touch and let us know how this approach works for you!
Speed up Automated Reasoning policy validation with conversational refinement ⚡👻🔍
Use Kiro CLI to ask questions about your policies, understand behavior & address issues. #AWS #AmazonBedrock #generativeAI @kirodotdev
👉 go.aws/4aetXan
ALT Use Kiro CLI to refine Automated Reasoning Policies
Stefano Buliani retweeted
Apply rigorous guardrails to help mitigate factual inaccuracies 🔒⚡🎯
With Automated Reasoning checks in #AmazonBedrock Guardrails, you can validate outputs against encoded business rules with results that provide mathematical certainty. #AWS
👉 go.aws/43oe7px
ALT Build reliable AI systems with Automated Reasoning on Amazon Bedrock: Part 1
Stefano Buliani retweeted
Can #generativeAI be powerful & provably verified? 🎯💪💡
Automated Reasoning checks in #AmazonBedrock Guardrails delivers up to 99% verification accuracy & provides auditable proofs for why the response is accurate. #responsibleAI
👉 go.aws/4ojgVwn
ALT With Automated Reasoning checks in Amazon Bedrock Guardrails, you can verify responses with up to 99% accuracy
Stefano Buliani retweeted
What makes machine-based reasoning so hard to get right? 💡🧠🎯
Amazon VP & Distinguished Scientist Byron Cook shares 3 top challenges & how #AWS is tackling them with Automated Reasoning checks in #AmazonBedrock Guardrails. #ResponsibleAI
👉 go.aws/4mZCJeX
Stefano Buliani retweeted
Create configurable safeguards with Automated Reasoning checks in #AmazonBedrock Guardrails 🚀🛠️🛡️
In this new hands-on workshop, learn to create policies, validate outputs & mitigate hallucinations with Automated Reasoning checks. #AWS #generativeAI
👉 go.aws/3K0xH45
Stefano Buliani retweeted
Automated Reasoning checks combine LLMs and SAT solvers to provide a combination of simplicity, flexibility, and mathematical certainty that's can't be achieved with either approach alone.
Can't wait to see what customers build with this.
We’ve reached a new milestone in AI safety. Today, @awscloud is launching Automated Reasoning checks in Amazon Bedrock Guardrails, bringing a new level of mathematical certainty to your AI outputs.
What makes this truly revolutionary is our two-stage approach: we use LLMs to translate text into logical formulas, then apply constraint solvers to mathematically verify compliance with defined policies. Through mathematical logic—not probabilistic estimates—we achieve up to 99% accuracy in identifying correct model responses.
We’ve used automated reasoning to secure AWS cloud infrastructure for more than a decade. Now, we're taking that proven approach and using it to make generative AI safer for customers.
Our partners at @PwC demonstrate how this industry-first technology is accelerating innovative ideas to market and delivering on stronger accuracy and compliance. More here: aws.amazon.com/blogs/aws/min…
Stefano Buliani retweeted
I’ll be on #AWSonAir today discussing how Automated Reasoning checks uses neurosymbolic framework to validate correctness in AI systems. Join us!
📺 Check out #AWSOnAir this Friday at 12 PM PT 👉 awsonair.net/AWSOnAir-Weekly
The show will discuss two new features:
🤖 See how Amazon #Bedrock verifies AI outputs with mathematical precision for accurate results
🚀 Learn how Amazon #Sagemaker HyperPod has faster model training
Stefano Buliani retweeted
1/3 AWS is tackling AI's biggest hurdle in regulated industries: trust. Their neurosymbolic AI on Bedrock, with automated reasoning checks, aims to deliver truly safe & explainable agent automation. This is key for enterprise adoption.
Hey @lyft just received an email that somebody removed my phone number from the account and the email has changed too, apparently. I cannot log in anymore. Please stop whoever is doing this
One of these @ktm_canada nights. Tomorrow I’ll get to the radiator and the damn leaking hose. Funny thing is that it seems to leak only when I don’t use the bike for a few days - does not leak the days I use it 🙃
So... no international wires with IBAN, cannot deposit USD checks (computer says no), 2 weeks of money unavailable converting currencies. It was easier to bank in Africa than in Canada. Is it @TD_Canada that sucks or is it banking in Canada in general that is stuck in the 1960s?