Rusnak Studio Ohad Rusnak
Menu
Skills Work Stack About Contact
Email

Available for hire · Contract & consulting · Remote-first · Worldwide

AI integration & production systems — architecture through deployment.

Senior engineer specializing in integrating AI into real products: multimodal pipelines, LLM-backed workflows, and scalable backends (FastAPI, Kafka). Ships systems that survive production — not demos.

Background spans edge vision (NVIDIA Jetson, real-time video), cloud inference, and mission-critical programs. I was principal architect and lead developer on Cargoseer — the customs risk platform later acquired by BigBear.ai. Open to senior IC roles, architecture consulting, and selective contract delivery.

Email for opportunities LinkedIn Selected work
  • FastAPI · Kafka · PostgreSQL · Docker (cloud & on-prem)
  • Computer vision, OCR, and LLM workflows in production
  • NVIDIA Jetson & GPU pipelines — edge to data center
Scroll

Snapshot

Key skills

  • AI & LLM integration

    Workflows, APIs, guardrails, production inference paths

  • Backend & platforms

    FastAPI, Kafka, PostgreSQL, Docker — AWS & GCP

  • Computer vision & OCR

    Multimodal pipelines, ONNX, inspection workflows

  • Real-time & edge

    Jetson, GPU video pipelines, latency-aware systems

  • Architecture & ownership

    System design, microservices boundaries, delivery

  • Quality & ops

    pytest, CI/CD, multiprocessing for throughput-critical paths

Proof of work

Selected projects

  • Principal architect and lead developer on Cargoseer (customs risk management); the product was a major factor in the company’s acquisition by BigBear.ai.
  • Took multimodal AI programs from concept through production deployment — including cloud and on-premises footprints.
  • Architected real-time inspection and vision workflows where latency, reliability, and field integration mattered.
  • Helped organizations turn complex AI ideas into operable products: clear interfaces, staged delivery, and testable milestones.
  • Worked on systems used in demanding, real-world environments — including defense-sector programs where failure modes are not acceptable.
  • Maintained a single accountable thread from design review to shipped artifacts, scaling with a small trusted network when bandwidth required it.

Case study — Risk management system for customs inspection

Cargoseer — customs risk management (X-ray & documents)

Principal architect & lead developer · Cargoseer Ltd. (now BigBear.ai) · customs and port operations

Challenge

Organizations need a coherent risk view across high-throughput cargo imaging and related documents — with reliability, auditability, and uptime suitable for live operations.

Execution

In the principal architect and lead developer role, I owned system architecture and core implementation: Python and FastAPI services, Kafka-driven microservices, PostgreSQL, Docker deployments on cloud and on-prem (AWS, GCP), and GPU-accelerated ONNX inference in data-center paths — embedded in full risk-assessment workflows.

Outcome

The platform became a major factor in BigBear.ai’s acquisition of Cargoseer. Delivered product lines (including Inspector and Cortex) that combine vision, OCR, and reasoning-style steps into operational customs risk workflows — including full-country deployment and integrations with external AI, databases, and models.

Shared Kafka-centric platform across multiple product surfaces Cloud and on-prem capable deployment model Multimodal production path: vision + OCR + LLM-style reasoning inside governed workflows

Stack

Tech stack

Tools and platforms I use regularly in production contexts.

  • Python
  • FastAPI
  • Apache Kafka
  • Microservices
  • PostgreSQL
  • Docker
  • AWS & GCP
  • ONNX
  • Computer vision
  • OCR pipelines
  • LLM integrations
  • Risk & decision workflows
  • Multiprocessing & concurrency
  • Testing (pytest, CI)
  • Linux
  • MongoDB
  • NVIDIA Jetson · embedded / edge
  • C / C++

Profile

About

Electrical engineer (Technion B.Sc.) turned hands-on AI/software lead. I was principal architect and lead developer on Cargoseer — the customs risk platform that became a major part of why BigBear.ai acquired the company. Broader work includes production risk and inspection systems for customs and defense-adjacent programs — from Kafka services to field-deployed vision stacks. I communicate clearly with product and research teams; full CV on request.

Education & background

B.Sc. Electrical Engineering, Technion — circuits & VLSI, control & robotics, signal & image processing. Dean’s List (three semesters). Graduation project: robotics, control, and 3-D vision with Kinect (featured in international press).

Languages: Hebrew & English — native · Spanish — basic.

Earlier: undergraduate mentor and lab technician, Control & Robotics Laboratory, Technion EE Faculty.

IDF Navy — missile-boat flotilla: weapon-system technician training, command course, commander of the electronic department (25 personnel), released 2009 (Staff Sergeant).

Remote-first, worldwide. Engagements favor clarity: defined outcomes, direct access, and documentation that survives handoff. The studio stays small by design; I bring in a trusted engineer network only when the program truly requires it.

  • Discovery: A short conversation to confirm fit, constraints, and whether a phased proposal makes sense.
  • Delivery: Build, integration, and hardening with visible milestones — pytest and CI where appropriate.
  • Enterprise & partners: Structured statements of work with product companies and integrators; senior IC depth throughout.
  • Scale when needed: Additional capacity through a vetted network — without diluting architectural coherence.

Contact

Let’s connect

Recruiters and hiring teams: share the role, stack, and timeline. Founders and leads: outline the problem space. I reply promptly to serious inquiries — full CV and references available on request.

Email [email protected] LinkedIn

Consulting rates $150/hr development · $200/hr architecture & advisory