I’m Dragutin Misiraca, founder of CoreX AI (DMCC, Dubai). I design and ship AI systems across healthcare, cybersecurity, crypto intelligence, and operations — multi-agent orchestration, retrieval-augmented generation, vector search, and production-grade evaluation loops — with a bias toward clear scope, fast delivery, and hardened deployment.
Most of the work lives where orchestration, product delivery, and security constraints all meet. These are the projects carrying the clearest signal right now.
A startup company for deployable AI systems. The work spans voice workflows, security research, crypto intelligence, internal automation, and product experiments that are designed to make it past demo day.
corexsoftware.aiTurns leases, regulatory filings, and market data into structured intelligence for real estate operators. Compliance monitoring, portfolio analytics, anti-scam detection, and automated reporting for boards and stakeholders.
Project detailsConversational scheduling and multilingual voice operations for clinics. Built around real call flows, latency constraints, and the practical messiness of patient-facing operations.
medcareflow.comCrypto intelligence platform built around specialized agents, retrieval, knowledge graph patterns, market feeds, and short briefings designed for faster signal review.
smartcryptoradar.aiA threat briefing and vulnerability intelligence project focused on grounded retrieval, operator-ready output, and the signal needed for security teams to act quickly.
Multilingual guest concierge with voice and chat, plus front desk and housekeeping workflow automation. Revenue analytics and sentiment tracking built on the same data layer.
Project detailsAI-powered digital forensics for macOS incidents. ForensiX ingests evidence, identifies file formats, reconstructs timelines, and flags anomalies before turning the case into clear, actionable intelligence.
Project detailsMobile check-in and check-out for Dubai hotels and premium hospitality brands. In-app service ordering, queue optimization, and deep PMS integration for real-time availability.
Project detailsAutomates web scraping, data collection, and monitoring across public, regulatory, and government sources. Continuous change detection and structured intelligence output, ready to drop into investment workflows.
Project detailsmacOS endpoint security designed to run entirely on Apple Silicon — zero telemetry, nothing sent to cloud. Six modules including a deception engine, a financial honeypot, and court-ready OSINT automation. Pending Apple ESF entitlement.
Project detailsDeveloper prompt optimization tool for Claude Code that turns rough implementation notes into structured, context-aware instructions with codebase awareness.
Real-time AI conversation for mental wellness support, with intelligent triage and specialist referral booking integrated into the flow. Encrypted channels, 24/7 availability, no appointment required.
Project detailsExecutive protection risk and route-planning concept with geospatial threat scoring, alternative route generation, and automated assessment reporting.
SIRA-certified personal protection and executive security in Dubai and Abu Dhabi. Discreet close protection trusted by HNW clients and families for private itineraries and event safety.
freakofnaturejames.comAI-guided arm specialization on iOS and Android — a structured 60-day program with pro bodybuilder video demos and PAR-Q+ safety screening built into onboarding. Four subscription tiers and a post-program retention path through maintenance cycles.
Project detailsThe best systems are usually not the loudest ones. I care more about whether something deploys cleanly, explains itself, and keeps behaving once real users get involved.
Small surface area, clear failure modes, measurable outputs. That matters more than a broad feature list.
I use retrieval, schemas, and evaluation loops to keep systems useful when the input is messy and the stakes are real.
The same rigor applied to CI, infrastructure, and browser automation should follow the AI layer too.
Before founding CoreX AI, most of my background was in quality engineering, browser automation, and security-heavy testing. That bias still shapes how I build.
Automated UI and backend testing with Java, Selenium, RestAssured, and TestNG, improving release coverage and helping reduce production downtime by 90%. Introduced Cypress, JavaScript, and Cucumber coverage for UI workflows. Covered Android and iOS flows with Appium.
Used JMeter to identify performance bottlenecks to support optimizations that increased system throughput by 50%. Created a C# and Selenium automation framework that reduced a one-day manual regression cycle to roughly 3 hours. Executed and maintained 100+ Appium scripts across iOS and Android.
Računarski fakultet. Bachelor with honours in electrical engineering and computing, focused on computer science and engineering.