Daniel Benassayag

AI Systems Engineer (Agents · RAG · Real-Time AI Platforms)

📍 Tekoa, Israel · 📞 058-433-5665 · ✉️ danielbenassaya2626@gmail.com
🇫🇷 Did aliya from france in 2011

Summary

AI Systems Engineer specializing in end-to-end intelligent systems, combining LLM agents, real-time computer vision, and distributed infrastructure.

Experience building:

  • multi-camera real-time AI surveillance systems (GPU-based)
  • RAG-powered production assistants
  • multi-tenant agent orchestration platforms

Strong focus on production-grade systems, not just models.

Experience

Lead AI Engineer — Ziv (IDF) (2025–2026)

Real-time multi-camera AI surveillance platform deployed across IDF kitchens.

  • Designed and deployed a real-time GPU inference pipeline processing multiple video streams for theft and kashrut violation detection
  • Built DeepStream-based CV system (person, hand, face detection + SigLIP embeddings + ReID clustering)
  • Developed event detection pipeline (moment detection → reranking → VLM validation → clip generation & redaction)
  • Architected backend services using FastAPI, Kafka, Redis for low-latency event processing and alerting
  • Implemented operator dashboard (React) with live camera feeds, identity tracking, and log streaming
  • Delivered end-to-end system: ingestion → inference → decision → alert → UI

Stack: Python, DeepStream (NVIDIA), Kafka, Redis, FastAPI, React, Docker, GPU

Founder / Lead Engineer — Agents24 (2026–Present)

Multi-tenant AI platform for building and deploying agent-based applications.

  • Built a multi-tenant AI control plane + execution platform for agents, RAG pipelines, and AI-powered apps
  • Designed graph-based agent orchestration system with compilation, streaming execution, and traceability
  • Implemented shared infrastructure for tools, artifacts, models, and knowledge stores
  • Developed RAG pipeline system with ingestion, retrieval, and operator-based workflows
  • Created AI coding agent + app builder runtime (sandboxed Vite apps per user)
  • Built governance layer: tenant isolation, RBAC, execution policies, observability

AI Engineer — Ravbot (IDF Military Rabbinate) (2024)

  • Built production RAG system answering religious and operational questions
  • Integrated official Military Rabbinate knowledge sources with retrieval pipelines
  • Designed uncertainty routing system → escalates low-confidence answers to human rabbis

Full Stack Developer — Natan (2021 · 6 months)

  • Developed features for medical clinic management system

Stack: .NET Core, Angular, MongoDB, Docker

Projects & Research

Reinforcement Learning Systems

  • Implemented DQN, REINFORCE, A2C, DDPG from scratch in PyTorch
  • Built custom 3D humanoid walking environment (PyBullet + DDPG)

Predictive Modeling

  • NBA game outcome predictor (~70% accuracy)

Low-Level ML

  • Built neural network framework from scratch (NumPy only)

OCR Research

  • Attempted Hebrew cursive OCR → identified dataset bottleneck

Education

B.Sc. Computer Science (partial) — Open University of Israel (2021–2024)

  • Linear Algebra I & II
  • Calculus I & II
  • Probability & Statistics
  • Discrete Mathematics
  • Data Structures & Algorithms

Magshimim Cyber Program (2019–2021)

  • Offensive & defensive security, OS internals, networking
  • Final project: full-stack social network with RSA/JWT auth + ML moderation

Skills

AI Systems
LLM Agents · RAG Pipelines · Reinforcement Learning · Transformers · Computer Vision

Backend & Systems
Python (FastAPI) · .NET Core · REST · WebSockets

Infra & Data
Docker · Kubernetes · Kafka · Redis · PostgreSQL · MongoDB · Celery · GPU pipelines

Frontend
React · Angular · Next.js

Security
Cryptography (RSA, AES, ECC) · CTF

Languages
French (native) · Hebrew (native) · English (fluent)