Oleks Bodryk (@obodryk) profile picture

- Cyber Porfolio Head for TOP-2 Global payments company xNetcracker, TNK-BP, iSIGHT Partners

Dubai
Joined May 2009
Photo by Oleks Bodryk (@obodryk)Photo by Oleks Bodryk (@obodryk)Photo by Oleks Bodryk (@obodryk)Photo by Oleks Bodryk (@obodryk)Photo by Oleks Bodryk (@obodryk)Photo by Oleks Bodryk (@obodryk)Photo by Oleks Bodryk (@obodryk)Photo by Oleks Bodryk (@obodryk)Photo by Oleks Bodryk (@obodryk)Photo by Oleks Bodryk (@obodryk)
Tokenmaxxing is interesting - you get to measure adoption and to fuel it generously at the initial stage, that said unchecked cost of AI would >> personnel cost, defeating the purpose for most use cases. North star though is scalable & predictable processes, not cost savings
13
Introducing the Open Knowledge Format (OKF), an open specification that formalizes the LLM-wiki pattern into a portable, interoperable format. AI is only as smart as the context we give it. As we build more advanced, agentic AI systems, they need accurate metadata and context to be useful. But in most organizations, that context is locked inside fragmented data catalogs, isolated wikis, scattered code comments, or the minds of senior engineers. Every time a new AI agent is built, teams are forced to solve the exact same context-assembly problem from scratch. To solve this, we've announced OKF, a vendor-neutral, open specification that formalizes the "LLM-wiki pattern" into a portable, interoperable format. It provides a standardized way to represent the enterprise knowledge that modern AI systems rely on. — Just markdown: readable in any editor, renderable on GitHub, indexable by any search tool — Just files: shippable as a tarball, hostable in any git repo, mountable on any filesystem — Just YAML frontmatter: for the small set of structured fields that need to be queryable: type, title, description, resource, tags, and timestamp We’ve also shipped reference implementations to help you hit the ground running, including an enrichment agent for BigQuery, a static HTML visualizer, and live sample bundles on @github → goo.gle/4uGvAEe ➕ Knowledge Catalog can now natively ingest OKF! Stop reinventing data models and building bespoke integrations for every new AI tool. Here's more about how OKF works → goo.gle/4uGvBbg
137
636
135
5,077
469,260
Oleks Bodryk retweeted
🚨 A Netflix engineer built an open-source proxy that cuts AI token usage by 60-95%. Zero code changes. Benchmarks show ±0.000 accuracy regression. ✨ 29.9k stars on GitHub. It sits between your app and the LLM, so every tool output, code block, and conversation history gets compressed in-flight. 🚫 No summarization, no loss. 😎 Just 60-95% fewer tokens with the same answers. Works with Claude Code, Cursor, Copilot, and any OpenAI-compatible client. One pip install, one env var, done. Netflix uses it internally. Apache 2.0. Built by Tejas Chopra. github.com/chopratejas/headr…
91
217
24
2,654
351,158
A little bit on AI transformation: identify meaningful business processes, see where AI help would be meaningful, feasible and viable, design and deploy on subset of employees, check benefits, keep or delete. Rinse & repeat
9
Any cybersecurity risk convo with business: 1) there is no risk; 2) there is no money to treat/avoid it; 3) its not clear what we are taking about. Outliers are usually ones when instead of cybersecurity risk a business risk is discussed, where cybersecurity is a backdrop only
11
Oleks Bodryk retweeted
754 structured cybersecurity skills for - AI agents 📍 Source: github.com/mukul975/Anthropi…
1
163
1
865
38,979
some AI specific breaches are highly likely, both because move fast and break things and because pushing new engineering heights eventually breaks production in a lot of unpredictable ways
Replying to @a16z
7x in tokens breaks per-app security. filters watching individual requests cannot scale to 19 billion tokens a minute. defense moves into the inference platform. that means scanning token sequences before the model responds and auditing every agentic tool call. 375 enterprises crossed 1T tokens last year. each runs a production workload where one bad prompt can leak customer data through a connected tool. platform-speed runtime guardrails do not exist as production-grade tools yet. someone builds that next.
55
Certainly, this is expanding TAM and utility, adding there - storage > big data > machine learning
unavailable
30
What is a "high severity vulnerability" for IT folks? That one, when they'll get calls at night and found themselves on an emergency meeting and see business, is accepting downtime and change implementation risk straight away #cybersecurity #vulnerabilitymanagement
we will win there will be new cities there will be new dreams there will be a new story there will be, there’s no doubt and those we've lost will be remembered and we will sing again, and we will celebrate anew 🇺🇦President ⁦@ZelenskyyUa⁩
1,666
11,370
1,315
52,211
In Memoriam: Yulia Zdanovskaya, a 21-year old mathematician, was killed on March 8th, 2022 during a Russian forces attack on Kharkiv, Ukraine. In 2017, Yulia represented Ukraine at the European Girls' Mathematical Olympiad and won a silver medal at the competition.
604
10,872
1,027
36,917
The one thing that security industry lacks more than everything - realistic assessment of its own capabiltiies
Even if you're in rush & fear of an incident response effort remember - no tool makes sense if there will be nobody to tune and get value from it
1