Real-Time Developer

Luxoft; Gdańsk; On-site; PLN 210,000 - 320,000; Full time

Job summary

Luxoft in Gdańsk, Poland, seeks a Real-Time Developer to lead Windows kernel‑mode data pipelines and ML runtime integration for a hardware‑assisted security platform. You will design low‑latency, high‑throughput systems handling CPU telemetry and secure IPC between kernel and user space, with production‑grade reliability and on‑site collaboration.

The role requires extensive C/C++ (and Rust) experience, deep kernel programming knowledge, and hands‑on work with ML inference runtimes and security

Qualifications

• Experience developing Windows kernel-mode drivers or related low-level Windows components.
• 4+ years in systems programming using C or C++; Rust relevant with strong low-level experience.
• Hands-on Windows kernel programming and kernel-to-user-space communication.
• Experience building real-time, high-throughput, or low-latency software.
• Knowledge of shared memory, ring buffers, lock-free queues, asynchronous I/O, or multi-threaded pipelines.
• Strong understanding of concurrency, memory management, and performance-sensitive design.
• Strong debugging and profiling across kernel-mode and user-space Windows components.
• Experience creating production-ready software with robust error handling and tests.
• Ability to analyze and reduce latency, CPU overhead, memory contention.

Responsibilities

• Design and implement high-throughput, low-latency data pipelines transferring CPU telemetry to ML runtime.
• Develop reliable kernel-mode to user-space communication mechanisms.
• Maintain real-time inference integration layer, including model loading and input preparation.
• Integrate ML runtimes such as ONNX Runtime, OpenVINO, or equivalents.
• Design stable SDK/API layer to deliver threat signals to security-software partners.
• Support API lifecycle, versioning, error handling, and documentation.
• Implement per-process and per-thread attribution of telemetry and results.
• Profile and optimize pipeline throughput, latency, memory usage, synchronization, and CPU overhead.
• Develop unit, integration, regression, load, and stress tests for the runtime pipeline.
• Collaborate with Lab and ML Engineers to replay telemetry and validate models.
• Diagnose issues across kernel-mode, user-space, runtimes, and interfaces.
• Document runtime architecture, communication protocols, API specifications, performance characteristics, integration procedures.

Skills

Windows kernel-mode

C/C++

Rust

Real-time

IPC mechanisms

Profiling

Windows debugging

Security integration

Education

University degree in CS/EE

Tools

Windows Driver Kit

WinDbg

ETW tracing

KMDF/UMDF

ONNX Runtime

OpenVINO

Job description

• AMD is building a hardware-assisted security platform that uses silicon-level Performance Monitoring Counters (PMCs) and on-chip machine learning to detect advanced endpoint threats, including ransomware, fileless malware, and cryptojacking, at the processor layer, below OS-based evasion.
• The platform collects CPU behavioral telemetry, classifies it through an ML inference engine, and exposes threat signals to security-software partners through a standardized API.
• The project covers the full engineering path from silicon telemetry and data generation through ML training and validation, real-time inference, lab qualification, and partner integration.
• The Real-Time Developer will lead the runtime and integration workstream, including the low-latency Windows data pipeline, integration with the ML inference runtime, and development of the SDK/API layer used by security-software partners.
• This is a Windows-focused role requiring regular on-site work in Gdańsk. Additional implementation details will be shared during the recruitment process in line with the applicable confidentiality requirements.

Responsibilities:

• Design and implement high-throughput, low-latency data pipelines that transfer CPU telemetry from a Windows kernel-mode component to a user-space ML inference runtime.
• Develop reliable communication mechanisms between Windows kernel-mode and user-space components.
• Develop and maintain the real-time inference integration layer, including model loading, input preparation, inference execution, result handling, and error recovery.
• Integrate production-compatible ML inference runtimes such as ONNX Runtime, OpenVINO, or equivalent technologies.
• Design and implement a stable SDK/API layer for delivering approved threat signals to security-software partners.
• Support API lifecycle concerns such as consumer registration, compatibility, versioning, error handling, and integration documentation.
• Implement accurate per-process and per-thread attribution of collected telemetry and inference results.
• Profile and optimize pipeline throughput, latency, memory use, synchronization, and CPU overhead.
• Design for concurrency and predictable performance using suitable mechanisms such as shared memory, ring buffers, lock-free queues, asynchronous processing, or equivalent techniques.
• Develop unit, integration, regression, load, and stress tests for the runtime pipeline and partner-integration layer.
• Support hardware-in-the-loop verification across relevant Windows platforms and system configurations.
• Collaborate with Lab and ML Engineers to replay captured telemetry, validate model integration, and investigate end-to-end behavior.
• Diagnose and resolve issues across kernel-mode components, user-space services, inference runtimes, and partner-facing interfaces.
• Document the runtime architecture, communication protocols, API specifications, performance characteristics, integration procedures, and known limitations.
• Work regularly on-site in Gdańsk and collaborate closely with the hardware validation environment.

Mandatory Skills Description:

• Experience developing Windows kernel-mode drivers and/or closely related low-level Windows components.
• 4+ years of professional experience in systems programming using C or C++. Rust experience is also relevant when supported by strong low-level systems experience.
• Hands-on experience with Windows kernel programming concepts and driver‑to‑user-space communication mechanisms.
• Experience building real‑time, near‑real‑time, high‑throughput, or low‑latency software.
• Practical knowledge of one or more high-performance communication or concurrency mechanisms, such as shared memory, memory‑mapped I/O, ring buffers, lock‑free queues, asynchronous I/O, or multi‑threaded pipelines.
• Strong understanding of concurrency, synchronization, memory management, process and thread behavior, and performance‑sensitive software design.
• Strong debugging and profiling skills across kernel-mode and user‑space Windows components.
• Experience developing reliable production‑oriented software with clear error handling, diagnostics, testing, and maintainability.
• Experience developing unit, integration, regression, load, or stress tests for systems software.
• Ability to analyze and reduce latency, CPU overhead, memory use, contention, and other runtime‑performance bottlenecks.
• Ability to work with technical specifications and document architecture, interfaces, protocols, and integration procedures.
• Ability to work regularly on-site in Gdańsk.
• University degree in computer science, electrical engineering, computer engineering, or an equivalent field.

Nice-to-Have Skills Description:

• Experience integrating ML inference runtimes such as ONNX Runtime, OpenVINO, TensorRT, DirectML, or equivalent technologies.
• Experience developing SDKs, libraries, or APIs used by external partners or multiple independent consumers.
• Experience with API versioning, backward compatibility, schema design, subscriber management, or partner‑integration support.
• Familiarity with x86-64 hardware performance monitoring, CPU telemetry, PMCs, PMUs, system tracing, or processor profiling.
• Experience with Windows Driver Frameworks, Windows Driver Kit, WinDbg, ETW/ETL tracing, KMDF, UMDF, IOCTLs, or related Windows technologies.
• Experience with hardware‑in‑the‑loop testing, platform validation, or integration across multiple hardware configurations.
• Experience optimizing software for GPU, NPU, DSP, or other hardware accelerators.
• Familiarity with AMD GPU compute technologies, Windows AI acceleration, or comparable GPU/NPU programming models.
• Experience with per‑process or per‑thread telemetry attribution.
• Background in endpoint security, EDR software, system monitoring, anti‑malware, or other security‑sensitive systems.
• Understanding of model input preparation, batching, inference execution, result interpretation, and runtime failure handling.
• Familiarity with driver signing or Windows driver certification processes. Prior WHQL experience is helpful but is not required for the Proof‑of‑Concept phase.
• Experience working across firmware, drivers, system services, hardware platforms, and application‑level interfaces.