Member of Technical Staff, AI @ HigherYou · Co-Founder & Security Engineer @ Aydo

I build production systems across applied AI/NLP, backend and data engineering, reliability, and Windows endpoint security.

• At HigherYou, I own applied AI systems end-to-end across data ingestion, information extraction, normalization, semantic/entity matching, ranking, validation, backend integration, evaluation, and production reliability.
• At Aydo, I co-build a Windows endpoint-protection platform spanning kernel/user-mode components, endpoint telemetry, static/dynamic analysis, C++ backend services, and isolated malware sandboxing.
• B.Sc. Computer Science @ Bar-Ilan University · GPA 89 · Expected Sep 2027

Windows Endpoint Protection Platform — EPP / AV / EDR

A Windows security platform spanning kernel/user-mode telemetry, endpoint services, static and dynamic detection, isolated malware analysis, backend infrastructure, and desktop management.

Highlights: Windows kernel driver + IOCTL · ETW/Sysmon telemetry · YARA/Sigma detection flows · VMware sandboxing · C++20/Drogon backend · PostgreSQL · installer/update/self-test infrastructure.

Evidence: Architecture · Security policy · CI

NLP Email Classification + Governed ML Data Pipeline

End-to-end NLP/ML engineering around reproducible and auditable data preparation, evaluation, and inference.

Highlights: source provenance and allowed-use metadata · data-quality reporting · exact/conflicting-label deduplication · template clustering · deterministic train/validation/test splits · explicit leakage checks · transformer training/inference · .eml and read-only Gmail workflows · CI-tested pipeline invariants.

Recorded experiment: final test F1 99.24%, ROC AUC 99.97% in the committed notebook run; documented with explicit provenance and generalization caveats.

Evidence: Results · ML engineering invariants · CI

PC-Free iOS App Refresh for Free Apple Accounts

An iOS transport and automation project that lets Free Apple Account users keep SideStore and two personally signed apps refreshed on the same iPhone, without a PC during refresh runtime.

Highlights: same-device app renewal · CoreDeviceProxy over TLS · contiguous CDTunnel framing · RSD · AFC staging · InstallationProxy · jktcp userspace TCP flow control · heartbeat lifecycle · native iOS background refresh scheduling.

Evidence: Architecture and build · Verification record · Release

Full-Stack Multi-Client Platform + C++ Recommendation System — Team Project

Collaborative food-delivery platform combining a C++ TCP recommendation service, Node.js / Express API, React, React Native / Expo, MongoDB, JWT, and Docker.

My contribution is visible in 12 merged PRs, including recommendation/persistence architecture, ranking logic, backend MVC/REST workflows, search, restaurant/menu management, mobile authentication, orders, and Docker integration.

Representative PRs: #12 — C++ recommendation/data architecture + 46 tests · #29 — restaurant/menu management · #42 — mobile auth · #43 — cart/orders/history · #44 — management + Docker

C++ Chess Engine + Windows IPC — Collaborative Project

Deterministic chess-engine logic with legal-move validation, check/checkmate/stalemate detection, transactional rollback for self-check prevention, board serialization, Windows named-pipe IPC, automated tests, CMake, and Linux/Windows CI.

Evidence: Modernization PR #1 · CI

Reliability-Focused Cross-Platform Monitoring

Playwright-based monitoring with a modular Python package separating browser integration, snapshot persistence, runtime orchestration, and reliability controls; built around persistent sessions, a persistent CLOSED / OPEN / HALF_OPEN circuit breaker, request/rate budgets, retry/backoff, Retry-After handling, challenge detection, failure-safe recovery, deterministic tests, and CI.

Evidence: Modular package · Reliability invariants · CI

Production engineering at HigherYou since 2023; Member of Technical Staff, AI since Dec 2025.

Work includes:

• NLP/ML extraction, classification, semantic/entity matching, ranking, validation, and decision pipelines
• large-scale/unstructured data ingestion and data-quality controls
• asynchronous backend services, APIs, WebSockets, PostgreSQL/SQLAlchemy, and external integrations
• ML evaluation: calibration, shadow/replay evaluation, statistical evidence gates, experiment tracking, and leakage controls
• multilingual speech/NLP pipelines including ASR, diarization, embeddings, and translation
• production reliability: observability, retries, health checks, diagnostics, retention, and failure recovery
• secure integrations with OAuth 2.0 / PKCE, JWT, encrypted secret storage, and audit logging

Product names, customer data, internal datasets, and private business logic are intentionally excluded.

• AntiVirus — hardened C signature scanner with recursive traversal, bounded binary matching, normal/quick modes, deterministic MSVC smoke tests, and CI; historical progression toward AYDO.
• TheLost — BLE proximity prototype using Android/Java, Python/Bleak, RSSI estimation, Firebase, tests, and CI.
• GifMaker — C frame-sequence editor using linked structures, persistence, and OpenCV.
• Shoot-It — contributed Windows implementation/fixes to a cross-platform screenshot automation utility.

Languages
C · C++ · Python · Java · JavaScript · SQL · x86 Assembly

AI / ML / NLP
PyTorch · Transformers · NLP · Information Extraction · Semantic Matching · Ranking · NER · Embeddings · ASR · Speaker Diarization · Model Evaluation · Calibration · Dataset Governance · Leakage Prevention

Systems & Security
Windows Internals · Kernel/User-Mode · EPP/EDR · ETW · IOCTL · Win32 API · Windows Services · Sysmon · YARA · Sigma · IPC · Reverse Engineering · Static/Dynamic Analysis · Malware Sandboxing

Backend & Data
FastAPI · Node.js · Express · PostgreSQL · MongoDB · SQLite · SQLAlchemy · Pydantic · REST · WebSockets · OAuth 2.0 / PKCE · JWT

Engineering
Docker · Nginx · CMake · GitHub Actions · Linux · PowerShell · Git · Testing · CI/CD · Observability · Structured Logging · Reliability Engineering

Problem → Architecture → Implementation → Testing → Evaluation → Deployment → Observability

I optimize for correctness, reliability, security, maintainability, data quality, measurable evaluation, and explicit failure behavior.