Jakub Atroszko

Backend systems, distributed performance analysis, benchmarking infrastructure, architectural simulation, and developer productivity tooling. • Co-invented a novel simulation acceleration methodology reducing architectural study turnaround by up to 194x while preserving high correlation to full-system simulation. • Created automated pre-silicon / post-silicon performance modeling frameworks integrating simulation, telemetry, and quality assessment pipelines. • Developed distributed systems performance analysis platforms spanning automated telemetry collection, large-scale data processing, and interactive analysis workflows. • Built internal tooling and full-stack applications used for performance debugging and optimization of modern cloud-native workloads. • Built benchmarking, observability, and telemetry infrastructure for industry-standard cloud/system benchmarks reflecting modern real-world workloads. • Delivered performance characterizations supporting major product launches and external performance claims. • Collaborated across organizational, technical, and geographical boundaries with senior stakeholders, engineering leadership, architects, and cross-functional teams to drive complex technical initiatives.

Distributed systems benchmarking and performance analysis of cloud-native workloads. • Delivered deep system-level and application-level performance analysis for Kubernetes-based distributed systems. • Investigated end-to-end latency, infrastructure bottlenecks, and hardware/software interaction across modern microservice workloads. • Produced actionable optimization recommendations based on telemetry, tracing, and performance profiling data.

Led and facilitated project-based and laboratory sessions for courses in Artificial Intelligence, Internet Services Architectures, and Parallel Processing on CUDA. • Taught AI concepts within a newly established interdisciplinary Data Engineering program for an international group of students. • Designed and delivered modern Python-based AI laboratory curriculum introduced as an industry-aligned alternative to legacy Matlab-based coursework, widely preferred by students and instructors. • Redesigned coursework around supervised learning, unsupervised learning, reinforcement learning, evolutionary algorithms, competitive learning - open-ended engineering problems to improve independent problem-solving and practical thinking.

Backend / infrastructure engineering contractor supporting internal data platform initiatives. • Developed infrastructure monitoring and observability solutions for on-premises environments. • Delivered backend functionality and automated tests for identity/access-management integrations. • Collaborated with stakeholders across teams to review and improve platform architecture and operational tooling.

Contributed across backend engineering, cloud infrastructure, data engineering, and operational tooling - working on large-scale automotive premium SaaS platform https://www.automotivemastermind.com/ and internal data workflows. • Automated previously manual business/product workflows using Microsoft SQL Server, Azure WebJobs, .NET Framework, and Microsoft Excel integrations, reducing turnaround time from days to hours while making the processes more reliable and less error-prone. • Delivered several end-to-end production features built on Microsoft Azure integrated with Google Cloud Platform (BigQuery, Kubernetes Engine, PubSub, Cloud Storage, Stackdriver) and third-party services, while designing and implementing monitoring, data analytics, and alerting infrastructure providing end-to-end coverage across all system components using Grafana, Prometheus, Weaveworks improving system visibility, reliability, and incident response.

Firmware validation, OS driver integration, and multi-team test infrastructure efforts.

Firmware and test infrastructure engineering for embedded / audio systems. • Built C++-based automated test infrastructure for firmware validation • Developed reusable testing framework, hardware mocks, integration tooling, and automation pipelines. • Performed low-level performance analysis and ISA comparisons for firmware/assembly-level optimization scenarios.

Interdisciplinary research at the intersection of artificial intelligence, cognitive science, signal processing, computational intelligence, and biological/computational system modeling. Published peer-reviewed work in artificial intelligence, human-in-the-loop AI, and signal classification.

Specialization: Distributed Applications and Internet Systems Thesis​ topic:​ (AI) Text​ classification​ with​ neural​ networks.

Engineering thesis: (AI) Comparison of ant colony algorithm, swarm algorithm and genetic algorithm in terms of exploration of objects in an intelligent two-dimensional space.