Zhihua Lai, PhD
| Senior Software Engineer |
Distributed Systems |
Platform & DevOps Engineering |
C/C++ |
Linux & Kubernetes |
Cloud & AI Infrastructure |
| 🏠Cambridge, UK |
📧 dr.zhihua.lai [AT] gmail.com |
đź“ž +44 (0)7939 518997 |
UK ILR |
PROFESSIONAL SUMMARY
- Senior Software Engineer at Microsoft Research Cambridge, working across AI/ML infrastructure, high-performance systems, robotics, cloud infrastructure, and 5G/6G networking
- Former AWS engineer and Day-1 contributor to Amazon S3 Object Lambda
- Former General Electric Staff Software Engineer, contributing to the Smallworld GIS platform and Magik programming language
- PhD in Computer Science, with research and publications in wireless systems and algorithms
- Top-3 Witness (Block Producer) on the Steem blockchain
Proven track record of delivering low-latency, production-grade systems across C/C++, Linux, eBPF, distributed cloud infrastructure, large-scale storage systems, and AI/ML platforms.
CORE SKILLS
- Systems & Networking: C, C++, Python, Linux, eBPF, low-latency systems
- Cloud & Distributed Systems: AWS, Azure, Kubernetes, cloud infrastructure
- AI/ML Infrastructure: GPU inference, ML tooling, compute platforms, robotics
- Research & Engineering: Algorithms, performance optimisation, rapid prototyping (0→1)
- Blockchain: Infrastructure, APIs, dApps, block production
PROFESSIONAL EXPERIENCE
Microsoft Research Cambridge — Senior Software Engineer
Intelligent Network Systems (formerly Azure for Operators)
https://www.microsoft.com/en-us/research/group/ins/
July 2021 – Present
- Contribute to AI tooling and standardised development images for Microsoft Research, enabling researchers to build, run, and experiment with AI workloads more efficiently.
- Build and improve the ML Proxy, a core infrastructure component for load balancing and routing inference requests across GPU backends, improving scalability and accelerator utilisation.
- Build components of the data flywheel and model-training pipeline for robotics applications, supporting continuous data ingestion, processing, training, and model improvement.
- Contribute extensively to a CLI for submitting, monitoring, and managing ML experiments across Microsoft compute platforms, simplifying researchers’ interaction with large-scale compute infrastructure.
- Contribute extensively to Project Janus, a userspace eBPF-based instrumentation and control framework for securely deploying monitoring and control functions.
- Developed Embedded Rust firmware on ESP32-C6 and integrated a robotic gripper with ROS2, bridging CAN-based robot commands to UART/RS485 hardware control and publishing device status back into the ROS2 stack.
- Evaluated power efficiency and throughput of ROS2 communication on Android, comparing a Kotlin bridge against native C++ CycloneDDS under controlled, equivalent traffic rates; implemented deadline-based pacing to prevent cumulative timing drift and ensure reproducible measurements.
I enable researchers to move rapidly from idea to working prototype by solving the engineering, infrastructure, and tooling problems around their research. My work spans eBPF in C, performance-critical C++, C#, Android, ROS2 robotics, and Embedded Rust on ESP32.
More recently, I have focused on AI/ML tooling at Microsoft Research, building systems that simplify experiment submission, execution, and infrastructure management across Microsoft compute platforms. This reduces operational friction, shortens research iteration cycles, and helps turn exploratory research into reproducible and scalable workflows.
Technical areas:
C, C++, Python, C#, Linux, eBPF, Azure, Kubernetes, AI/ML Infrastructure, GPU Inference, Compute Platforms, GitHub CI/CD
Selected Microsoft Open-Source Contributions
GitHub username: doctorlai-msrc
Hackathons
- 2026 — Spirit of Hackathon Award
- 2025 — Societal Impact Award
Selected Publications
Previous Publications
- Radio Propagation Modelling, Chapter 2 in Heterogeneous Cellular Networks: Theory, Simulation and Deployment, Cambridge University Press, 2012
- Outdoor-Indoor Channel, Chapter 5 in LTE-Advanced and Next Generation Wireless Networks: Channel Modelling and Propagation, John Wiley & Sons, 2012
- Intelligent Ray Launching Algorithm for Indoor Scenarios — Radioengineering / Towards EUCAP 2012
Amazon Web Services — Software Development Engineer (L5)
AWS S3, Cambridge, UK
January 2020 – July 2021
Core contributor to Amazon S3 Object Lambda.
- Contributed to the design, development, and global launch of Amazon S3 Object Lambda, a new AWS S3 capability serving customers at global cloud scale.
https://aws.amazon.com/s3/features/object-lambda/
- Built internal performance-analysis tooling to identify system bottlenecks, investigate latency, and improve operational visibility across highly distributed storage infrastructure.
- Supported production operations through an on-call rotation for business-critical AWS services, including S3 Object Lambda, troubleshooting live issues and maintaining service reliability.
- Worked across service development, infrastructure, diagnostics, performance analysis, and operational readiness in a large-scale distributed-systems environment.
Technical areas:
Java, Amazon S3, AWS Lambda, Distributed Systems, Performance Analysis, Cloud Infrastructure, Production Operations, On-Call
General Electric — Staff Software Engineer
Smallworld GIS, Cambridge, UK
December 2018 – January 2020
Core contributor to the Smallworld GIS platform and Magik programming language.
- Identified missing language capabilities in Magik and independently designed and implemented native
while loop and regular-expression support.
- Took both features from prototype through technical review, implementation, testing, backward-compatibility validation, and product approval.
- Worked across engineering and product teams to safely introduce language changes into a platform with millions of lines of production code.
- Both capabilities were accepted and shipped as part of the Smallworld 5.x release.
https://en.wikipedia.org/wiki/Magik_(programming_language)
Technical areas:
C, C++, Java, Magik, Programming Languages, Compiler/Language Development, Testing Frameworks, CI/CD
Ranplan Wireless — Principal Algorithm Engineer
Cambridgeshire, UK
September 2010 – November 2018
https://ranplanwireless.com/
Led the research, design, and implementation of radio-propagation algorithms for Ranplan’s Radio Propagation Model (RPM), one of the company’s core intellectual-property assets and a key component of its commercial wireless-network planning products.
- Lead developer of Ranplan’s radio-wave propagation engine, one of the company’s core commercial IP platforms.
- Designed and implemented ray-tracing, diffraction, and wireless channel-modelling algorithms used in commercial indoor and outdoor network-planning tools.
- Led a major C++ rewrite and performance optimisation of the propagation engine, including OpenCL GPU acceleration and performance-critical inline assembly.
- Researched and prototyped machine-learning algorithms for wireless modelling, e.g. KNN-based geolocation.
- Developed high-performance 32-bit and 64-bit propagation components and extensibility APIs using C++, Delphi, C#, and low-level optimisation techniques.
- Designed algorithms for processing and converting GIS, clutter, terrain, and building data into representations used by propagation simulations.
- Led development of supporting products and components including GeoConverter and GLS licensing services.
- Designed and implemented 1,500+ Win32/64 COM APIs in Delphi and built extensive unit and integration test suites.
- Developed internal engineering tooling for performance analysis, statistics, code quality, CI, and automated validation.
- Provided Level 3 technical support and customer consulting for complex wireless-planning and propagation-modelling scenarios.
- Contributed to Ranplan patents and technical research in radio propagation, channel modelling, and indoor positioning.
Technical areas:
C++, C#, Delphi, OpenCL, GPU Computing, Ray Tracing, Radio Propagation, Wireless Channel Modelling, GIS, Machine Learning, KNN, Performance Optimisation, CI/CD, Testing
EDUCATION
PhD, Computer Science — Wireless & Algorithms
University of Bedfordshire, UK · 2006–2010
Thesis: The Development of an Intelligent Ray Launching Algorithm for Wireless Network Planning
BSc, Computer Science — First Class Honours
University of Luton, UK · 2004–2006
Final Project: Chinese Chess
PATENTS — SELECTED
- Method for Predicting Outdoor 3D Signal Field Strength — PCT/GB2015/053224
- Method for Predicting Indoor 3D Signal Field Strength — PCT/GB2015/053223
- WiFi Multi-Band Fingerprint-Based Indoor Positioning — WO/2018/167500
BLOCKCHAIN & OPEN-SOURCE LEADERSHIP
- Top-3 Witness (Block Producer) on the Steem blockchain
- Architect and operator of Steem infrastructure, APIs, explorers, and services
- Maintainer of open-source blockchain tooling and load-balancing infrastructure
HONOURS & AWARDS
SELECTED MICROSOFT CERTIFICATIONS
Additional Microsoft certifications include Azure, AI, security, GitHub, Power Platform, and data credentials.
Full list of 28 Microsoft certifications:
https://doctorlai.github.io/microsoft-certifications
GOOGLE CLOUD CERTIFICATION
MISCELLANEOUS
