←
Codified HQ31 posts
Codified HQ avatar

Codified HQ

@CodifiedHq

AI Governance

◷ Joined May 2023▧ 8 Photos and videos
Codified HQ avatar
Codified HQ@CodifiedHq·14 Jan 2025

The Enterprise AI Guardrail that you always wanted but never had, DM us for more.

Yatharth@y9914 Jan 2025

Hey CIO & CISOs, .. there is no easier way to be safe & onboard AI tools in your Enterprise. 1⃣ A simple API call for the Gen AI App 2⃣ A fully managed SaaS interface for you to control who gets to see what & we make this happen for most customers in < 10 days. Reach out to us at codified.io #RAG #Cybersecurity #CISO #CIO #Governance #Data #EnterpriseAI

Post attachment
Codified HQ avatar
Codified HQ@CodifiedHq·13 Dec 2024

What could go wrong? 😎 . . . . . . . . . 🚀 AI is powerful, but trust is non-negotiable. - Do you really know what data your Gen AI is using to answer questions for your employees? - Does it securely differentiate between employees or just take a "one-size-fits-all" approach?

Post attachment
Yatharth avatar
⟳ Codified HQ retweeted
Yatharth@y99·18 Oct 2024

Arvind is pointing out something that we at @CodifiedHq have been observing as a common trend for AI adoption. 🟢 AI is making it easier for users to find what they need. 🔴 Enterprises are discovering that employees have access to things they shouldn't. This is a big problem. 🔴 Causes slow adoption, which impacts the rate of innovation. The Fix? We need a shift from RBAC to high-level, policy-driven access control. 1️⃣ Write policies for PII/PHI and classified data, not project-level RBAC. 2️⃣ Dynamic Access – eliminate unused access and reduce risks. 3️⃣ Build for intended use – systems learn and create safer policies over time. #datagovernance #dataaccess #accessgovernance #aigovernance

Security, governance, and AI safety in GenAI deployments are top of mind for CIOs. In fact, in a survey @glean conducted earlier this year, 73% of CIOs said that generative AI tools not vetted by the IT department are a threat to the business. AI security is a core focus for Glean and our customers. Enterprises rely on Glean for its strict permissions enforcement, single-tenant model, and AI governance practices. I shared some ways we approach AI security with @srimuppidi from @BusinessInsider that you can read in the article below, and also expanding on them here: 1️⃣ Permissions: LLMs should only access data that the user is allowed to see. This way, we keep sensitive information safe and avoid any data leaks. 2️⃣ Training: Since LLMs can remember the data they were trained on, it's crucial to never use company data for training. This helps protect proprietary company information. Glean has agreements with leading LLM providers that promise zero-day retention and no training on your data. 3️⃣ Evaluation: Our main focus is on work AI, so users expect answers that are based on company data. We use our own AI evaluators to ensure the answers generated by LLMs are accurate and reliable.

Ali Ghodsi avatar
⟳ Codified HQ retweeted
Ali Ghodsi@alighodsi·27 Mar 2024

Today we released an open source model, DBRX, that beats all previous open source models on the standard benchmarks. The model itself is a Mixture of Experts (MoE), that's roughly twice the brains (132B) but half the cost (36B) of Llama2-70B. Making it both smart and cheap. Since only 36B expert parameters are used live, it's close to twice the speed (tokens/seconds) of Llama2-70B. We're excited to build custom versions of this for organizations that have proprietary data! Check it out! databricks.com/blog/announci…

Madrona Venture Labs avatar
⟳ Codified HQ retweeted
Madrona Venture Labs@MadronaVL·19 Mar 2024

Codified CEO @y99 on why he chose to partner with @MadronaVL: "I thought MVL would compliment me in areas where I didn't have experience, and that expectation was more than met. And if you're looking at outcomes, which is the right way to evaluate this, what MVL did is accelerate the journey. It was not just about starting the company faster; it also accelerated us in weeding out bad ideas because we could do it together. It accelerated the hiring of people because I had help from day one. It accelerated building from concept to design. It accelerated, most importantly, conversations with customers. For founders, time is an incredibly valuable commodity. And that's what I got in partnering with MVL." Read more about his journey here: bit.ly/4bZfvBy #VentureCapitalMyths #DataInnovation #TechLeadership #DataGovernance #StartupFounder