HomePočetnaAboutO namaServicesUslugeISO 42001ISO 42001TeamTimIndustriesIndustrijeProjectsProjektiBlogBlogGet in TouchKontakt

Blog

Notes from our work in AI, blockchain, cybersecurity, and fintech.

· Company News · Novosti

Introducing the ResolverSys BlogPredstavljamo ResolverSys blog

Why we started writing - and what to expect here about AI, blockchain, cybersecurity, and the engineering behind our work.Zašto smo počeli da pišemo - i šta ovdje očekivati o AI, blockchainu, kibernetičkoj sigurnosti i inženjerskom radu iza naših projekata.

Read more →

· AI, LLM, AI Security · AI, LLM, AI sigurnost

TypeSafe Jev - A Model That Returns Decisions, Not ParagraphsTypeSafe Jev - model koji vraća odluke, a ne pasuse

Jev takes unstructured state and returns typed, calibrated probabilities that ordinary code can branch on. Here is how it works, where it stops, what a call costs, and how we are testing it inside an AI firewall.Jev prima nestrukturirano stanje i vraća tipizirane, kalibrisane vjerovatnoće po kojima običan kod može da se grana. Kako radi, gdje mu je granica, koliko košta jedan poziv i kako ga testiramo u AI firewallu.

Read more →

· Engineering, Delivery, Security Integration · Inženjering, Isporuka, Sigurnosne integracije

Forward Deployed Engineering - Building Software Where the Problem LivesForward Deployed Engineering - softver se pravi tamo gdje je problem

A forward deployed engineer owns a customer outcome from discovery to deployment and then turns field lessons into product. Here is what the role is, how an engagement runs day by day, and where it goes wrong.Forward deployed inženjer vodi rezultat za klijenta od otkrivanja problema do produkcije, a zatim lekcije sa terena pretvara u proizvod. Šta je ta uloga, kako izgleda angažman dan po dan i gdje sve može da pođe po zlu.

Read more →

· DevOps, Cloud Security, Infrastructure · DevOps, Cloud sigurnost, Infrastruktura

Smaller Attack Surface by Design - One Gateway, Private Backends, Zero-Touch TLSManja izloženost ugrađena u arhitekturu - jedan gateway, privatni backendi, TLS bez ručnog rada

Public IPs on every app server are convenient until they are not. One gateway, private backends, and automated TLS shrink what the internet can see.Javna IP adresa na svakom serveru djeluje praktično - dok ne postane problem. Jedan gateway, privatni backendi i automatski TLS smanjuju ono što internet uopšte može da vidi.

Read more →

· Consulting, Decision Making, AHP · Konsalting, Donošenje odluka, AHP

Structured Decision-Making - From Gut Feel to Defensible Choices with AHPStrukturirano donošenje odluka - od osjećaja do utemeljenih izbora sa AHP-om

Every engineering choice is a trade-off. The Pugh matrix screens options fast, but equal weights hide what actually matters. Analytic Hierarchy Process (AHP) turns pairwise comparisons into objective weights you can defend.Svaki inženjerski izbor je kompromis. Pughova matrica brzo sužava listu, ali jednake težine skrivaju ono što je zaista važno. Analitički hijerarhijski proces (AHP) pretvara poređenja u parovima u objektivne težine koje možete braniti.

Read more →

· Cybersecurity, TLS, Privacy · Kibernetička sigurnost, TLS, Privatnost

TLS 1.3, the SNI Paradox, and the Road to Encrypted Client HelloTLS 1.3, SNI paradoks i put ka Encrypted Client Hello

TLS 1.3 cut handshake latency in half and made forward secrecy mandatory - but Server Name Indication still leaks in plaintext. Encrypted Client Hello (ECH) is the fix, and adoption is moving slower than the spec.TLS 1.3 prepolovio je latenciju handshake-a i učinio forward secrecy obaveznim - ali Server Name Indication i dalje curi u otvorenom tekstu. Encrypted Client Hello (ECH) je rješenje, a usvajanje ide sporije od same specifikacije.

Read more →

· Cybersecurity, Zeek, Network Security · Kibernetička sigurnost, Zeek, Mrežna sigurnost

Mastering Zeek Scripting - Network Security Telemetry at ScaleOvladavanje Zeek skriptama - sigurnosna telemetrija mreže u velikom obimu

Zeek is a Turing-complete, event-driven language for decoding network protocols and generating security telemetry. Here is how native types, radix-tree lookups, and custom scripts like AppProfiler turn live traffic into actionable alerts.Zeek je Turing-kompletan, event-driven jezik za dekodiranje mrežnih protokola i generisanje sigurnosne telemetrije. Evo kako nativni tipovi, radix-tree pretrage i prilagođene skripte poput AppProfiler-a pretvaraju saobraćaj u upozorenja.

Read more →

· AI, SEO, GEO · AI, SEO, GEO

Generative Engine Optimization - Why SEO Alone Is No Longer EnoughGenerative Engine Optimization - zašto SEO više nije dovoljan

AI Overviews and LLMs are answering searches without clicks. GEO optimizes your content to be cited as a source in AI-generated answers - here is what that means in practice.AI Overviews i veliki jezički modeli odgovaraju na pretrage bez klika. GEO optimizuje sadržaj da bude citiran kao izvor u AI odgovorima - evo šta to znači u praksi.

Read more →

· AI, Cloud Security · AI, Cloud sigurnost

Why the Classic Cloud Is Unsafe for AI - and How TEE Solves ItZašto je klasičan cloud nesiguran za AI i kako TEE rješava taj problem

In the traditional cloud, your AI models, prompts, and data are exposed to the provider, the hypervisor, and root users. TEE and NVIDIA Confidential Computing bring hardware isolation - here is how they work.U tradicionalnom cloudu AI modeli, promptovi i podaci izloženi su provajderu, hipervizoru i root korisnicima. TEE i NVIDIA Confidential Computing donose hardversku izolaciju - evo kako funkcionišu.

Read more →