🎓 B.Sc. Software Engineering student at Afeka College (Academic Reserve – Atuda)

🔐 Focused on network security, binary exploitation and systems programming

💡 I build clean, practical, and maintainable software, and I test and benchmark what I build

Languages

Frameworks & Tools

• Low-Level, Binary Exploitation & Reverse Engineering: Linux ELF Binaries (32/64-bit), Memory Safety & Corruption (Buffer Overflows), Dynamic Debugging (GDB + GEF), Mitigation Analysis (checksec: Canary, NX, PIE), Static Analysis (Ghidra), Exploit Automation (pwntools, C, Python).

• Networking & Security Observability: OSI & TCP/IP (L2–L7), Deep Packet Inspection (DPI), Raw Sockets (Scapy), Traffic Analysis & Intrusion Detection (DoS, Port & Stealth Scans, DNS Tunneling), Active Defense (time-limited firewall blocks), Infrastructure-as-Code (Grafana, Loki, Promtail).

• Cryptography & Vault Architecture: Symmetric Encryption (AES-128-CBC / Fernet, HMAC-SHA256), Key Derivation (PBKDF2), Salting, Secure Vault Design.

• Backend Systems & Concurrency: Multi-threaded Architecture (Producer-Consumer Pipelines), Thread Synchronization (threading.Lock), C-FFI / Native Libraries (ctypes).

• DevOps, Testing & Infrastructure: CI/CD (GitHub Actions), Containerization (Docker), Automated Testing (pytest), Memory-Safety Fuzzing (ASan / UBSan), Infrastructure Automation & Scripting (Bash), PyPI Packaging.

Real-Time Network Intrusion Detection & Prevention with Dual-Layer DPI & a Code-Managed Observability Stack

• Motivation: Building a real-time NIDS/NIPS to monitor L3-L7 traffic, detect attacks, and respond through the OS firewall, while learning how network defense tools work internally.
• The Challenge: Scanning raw binary traffic quickly and safely: Python is slow per packet, the C engine must never read past a buffer (Memory Safety), and the system must avoid false positives (SYN-ACK replies, outbound traffic, encrypted payloads) and never block innocent hosts — including diagnosing and fixing a false-positive signature match against encrypted traffic during live testing.
• Key Learning: Implemented a dual-layer live DPI pipeline — an Aho-Corasick automaton for cleartext credentials alongside a Native C DPI Engine (ctypes) with 14 case-insensitive injection signatures and explicit bounds checks, scanning about 1M packets/s per call (about 5x a naive Python loop, up to 1.5x Aho-Corasick on larger payloads). Profiling showed that Scapy dissection accounts for about 80% of per-packet time while all DPI is under 3%. Audited my own code and fixed 8 defects, each pinned by a regression test; 46 pytest tests, C engine fuzzed under ASan/UBSan, GitHub Actions CI. Applied producer-consumer queues, thread safety (threading.Lock), time-limited firewall blocks, and Dashboards-as-Code.
• Tech Stack: Python, C (GCC/ctypes), Scapy, pytest, GitHub Actions, Multithreading, Docker Compose, Grafana, Loki, Promtail, JSON Structured Logging, IaC.

Personal security research repository with picoCTF binary-exploitation writeups, working exploit scripts, and mitigation analysis

• Motivation: Deepening practical skills in low-level vulnerabilities, memory safety, and binary analysis, one documented challenge at a time.
• The Challenge: Analyzing compiled ELF binaries, identifying memory-corruption flaws (stack buffer overflows, function-pointer misuse, unchecked gets() / strcpy()), and crafting exploits that redirect execution flow or change program logic.
• Key Learning: Four solved challenges covering return-address overwrite on a 32-bit binary (offsets found with De Bruijn patterns in GDB + GEF), function-pointer redirection through symbol lookup, and targeted stack-variable overwrite. Read the decompiled code in Ghidra to understand each binary, did the exploit development itself in GDB + GEF, read checksec output to see which mitigations (Canary, NX, PIE) actually matter, and automated every exploit with pwntools against local and remote targets.
• Status: Ongoing. Next platform: ROP Emporium.
• Tech Stack & Tools: C, Python, pwntools, GDB (GEF), Ghidra, checksec, Linux (WSL2/Ubuntu), Bash.

Secure, modular CLI/TUI password manager published on PyPI with automated testing and releases

• Motivation: Moving away from third-party managers to build a personal vault on vetted cryptographic primitives, and to learn end-to-end Python packaging and release automation.
• The Challenge: Correctly using Fernet authenticated encryption (AES-128-CBC & HMAC-SHA256) and PBKDF2 key derivation (600,000 iterations, per OWASP guidance) with a per-vault salt, so the master password is never stored; plus storage that survives crashes and an interactive arrow-key TUI (questionary/rich).
• Key Learning: Auditing my own security tool and fixing what I found: atomic vault writes, owner-only file permissions, a self-clearing clipboard and an upgrade path for older vaults. Separated the cryptographic core from the CLI/TUI, covered it with 54 pytest tests (76% coverage) running in CI on Python 3.10–3.13, and publish releases through PyPI Trusted Publishing (OIDC, signed attestations, approval-gated environment).
• Tech Stack: Python, Cryptography, Pyperclip, Rich, Questionary, Pytest, GitHub Actions (CI/CD, Trusted Publishing), PyPI.

🤝 Let's Connect! Whether it's about a project or a potential opportunity, I'm just a message away.