Computer Engineering student and co-founder of Interviewlary building applied-AI products, voice systems, agent workflows, full-stack applications, and developer infrastructure. Practical work spans product definition, repository-level AI-assisted implementation, technical review, integration, debugging, QA, deployment, and operations. The portfolio includes shipped public tools, live web products, and substantial systems under active development.
Co-founded an AI interview-training SaaS that generates role-specific questions, runs voice/text mock interviews, transcribes answers, and returns scoring, strengths, weaknesses, and actionable feedback.
Lead product planning, requirements, user journeys, prioritization, provider evaluation, pricing and positioning work, launch coordination, and operational decisions across the product lifecycle.
Perform technical review and QA across question generation, STT/TTS voice flows, scoring and feedback, authentication, subscription and billing flows, backend behavior, latency, and production edge cases.
Contribute through AI-assisted engineering workflows for repository inspection, implementation, integration, debugging, testing, and release iteration using Git/GitHub and modern coding agents.
Lead product communication and marketing direction, including launch materials, social content, demonstrations, outreach, and founder-facing operations.
Education
Istanbul Medipol University — Bachelor’s in Computer Engineering
Relevant academic areas: Data Structures & Algorithms, Web Programming, Computer Organization, Programming for Engineers, and Project Management.
Technical stack & applied areas
Programming
TypeScript, JavaScript, Python; academic exposure to C and Java.
LLM applications, AI agents, multi-agent workflows, RAG and evidence systems, STT/TTS voice AI, provider routing and fallbacks, semantic caching, MCP.
Engineering workflow
Git/GitHub, unit and integration testing, build/lint/typecheck gates, Sentry and observability, API/provider evaluation, AI-assisted development with Cursor, Codex, Gemini/Antigravity, and OpenCode.
Product & delivery
Product planning, business analysis, technical QA and review, user-flow analysis, provider evaluation, marketing and positioning, operational coordination.
Comprehensive Digital SAT preparation platform with a Socratic AI tutoring engine (server/tutor.ts) designed to guide reasoning without solution leakage, interactive knowledge map mastery trees, Three.js 3D mascot, 64-question curated curriculum, and Playwright cross-browser test suites.
TypeScript and Node optimization layer for multi-provider AI agents with exact and semantic caching, prompt optimization, real-dollar budget controls, provider adapters, and MCP tooling. 92/92 tests passing; 35.5% token reduction in repository mock A/B scenarios with identical output fidelity.
TypeScript · Node.js · SQLite · MCP
TruthNet
Award Winner · AI Hackathon
Real-time claim-level fact-checking system using an adversarial four-agent pipeline (Research, Prosecution, Defense, Judge) with evidence retrieval and live streaming. Built and awarded 1st place during a seven-hour hackathon sprint.
FastAPI · React · Gemini · SSE · Docker
Web Dev Skills
Shipped · Public
Research-backed AI-agent engineering playbook for production websites: 25 specialized skills, 41 lifecycle prompts, 15 product archetypes, agent instruction templates, and evidence, security, and release gates.
Agent workflows · Open source · Engineering process
Reliability Guard
Shipped · Public
Two-layer reliability framework for AI coding agents covering anti-fabrication rules, uncertainty, research verification, source conflicts, and evidence handling; includes 24 test scenarios and multi-agent compatibility.
AI agents · Evaluation · Reliability
Live products & shipped systems
RoutineAI
Shipped PWA · Active
Student day OS with conversational setup, deterministic scheduling, dashboard timeline, miss recovery, streaks, focus tools, alarms, and offline IndexedDB state. Repository gate passed with 106/106 tests plus build, lint, typecheck, and secret scan.
Next.js · React · TypeScript · Supabase · Dexie
StudyRoot AI
Shipped · Academic OS
Semester-long multi-course academic operating system supporting 7 active courses, 12 template types, embedded MS Office lecture previews, and synchronized YouTube timestamped note-taking.
Vite · React 18 · Supabase · Vercel Edge
Syria Trip
Shipped · Live
Phone-first bilingual guide for Damascus and Rif Dimashq with 100 curated places, Arabic and English search, filters, favorites, visited state, RTL support, and locally persisted preferences.
HTML · CSS · JavaScript · Bilingual UX · RTL
AI Courses & Guides
Shipped · Live
Bilingual interactive learning portal covering practical AI fundamentals, prompt engineering, agents, RAG, AI development tools, programming tracks, and Interviewlary practice guides.
HTML · JavaScript · Bilingual learning
StudyK Group
Shipped · Live
Education consultancy and lead-generation product with structured university data, AI admissions assistant, protected Supabase Edge Functions, rate limiting, and application/contact flows.
React · TypeScript · Supabase · AI assistant
KaisLearnAI
Live preview · In development
Production rebuild of a bilingual Arabic and English AI-learning platform. Foundation, catalog and starter path, auth and RLS migrations, protected learning contracts, motivation and rewards, and privacy and abuse controls are implemented locally; public preview is live.
Next.js · TypeScript · Supabase · RTL
Active engineering builds & architecture
Jadwalna AI
Active Prototype
Smart itinerary planning engine for Saudi Arabia featuring an Expo 54 cross-platform client, Fastify 5.2 constraint solver, Next.js 16 sharing, 1,000+ OSM Riyadh venues, and Saudi dialect NLP rules.
Expo 54 · Fastify 5.2 · Next.js 16 · TypeScript
Uniapy
Active Build
Bilingual university admissions & consultant CRM with Next.js 15.2, Prisma 6.5, Jose JWT RBAC, dynamic student profile extraction, and AR/EN RTL triage workflows.
Next.js 15.2 · Prisma 6.5 · Jose JWT · RTL
Lary OS
In development
Interviewlary founder command center aggregating AI briefings, finance, technical and provider monitoring, signed product reporting, scheduled email digests, and audited operational actions. Releases 0–4 implemented; Cloudflare Workers deployment architecture.
Next.js · TypeScript · Cloudflare Workers · AI agents
MCOS — Marketing & Content Intelligence OS
In development
Local-first research-to-marketing platform with workspace context, source and evidence handling, stale-fact and contradiction detection, trend scoring, content engines, bilingual carousel rendering, and approval gates. Verified host state includes 38 Python tests and a production Next.js build.
Python and FastAPI voice AI assistant with wake word, streaming Deepgram STT, Groq model routing, local TTS, Supabase memory, and Gmail, Telegram, Google Calendar, and MCP integrations.
Python · FastAPI · Voice AI · Supabase · MCP
Parallel System
Architecture Spec
Vendor-neutral control-plane architecture isolating autonomous coding agents in temporary Git worktrees with task routing, validation gates, and human approval checkpoints.
3D and time zebrafish cell-tracking competition. Submitted a three-model 3D U-Net ensemble with test-time augmentation, velocity-aware assignment, and graph repair; public score 0.917 and rank #445 at the captured point.
Python · 3D U-Net · Computer vision
Medipol META AI Hackathon
May 2026
Built and presented TruthNet with a team in a seven-hour sprint, taking the idea through agent pipeline, backend, interface, debugging, and a live stage demo.
AI agents · FastAPI · React
BirdCLEF+ 2026
Kaggle · 2026
Machine-learning competition work on identifying wildlife species from large-scale bioacoustic recordings.