Senior C++ Robotics & GPU Engineer

ANAND RATHI

Australian Citizen C++**/** • ROS* • CUDA/GPU • Distributed Systems

**********.***@*****.*** +61-402-***-*** Perth, WA

PROFESSIONAL SUMMARY

Senior C++ engineer, 20+ years, specializing in robotics, autonomous systems, GPU-accelerated perception, and

distributed systems architecture. RECENT: ROS2 Humble production work (robot control API, DDS bridging,

industrial protocol integration). Worked on : C++20 coroutines, Boost ASIO async I/O, CUDA inference pipelines,

TensorFlow/Keras ML frameworks, computer vision (YOLO, UNet semantic segmentation). TRACK RECORD: Solo full-stack

delivery—from algorithm design to system architecture to production deployment. TDD discipline. Comfortable

troubleshooting complex multi-threaded systems and commissioning in field environments.

TECHNICAL COMPETENCIES

ROBOTICS & AUTONOMOUS SYSTEMS

• ROS2 Humble (production): robot control API migration, DDS middleware, WebSocket bridges

• C++20 coroutines, Boost ASIO & Beast (async I/O networking)

• Industrial protocols: Modbus, EthernetIP, CAN, serial communication

• Multi-threaded event loops, real-time control constraints

• State estimation & sensor fusion (Kalman filter, LSTM)

GPU & MACHINE LEARNING PERCEPTION

• CUDA (NVIDIA GPU compute), TensorRT, ONNX inference optimization

• ML frameworks: TensorFlow, Keras, PyTorch

• Computer vision: YOLO object detection, UNet semantic segmentation, OpenCV

• Image processing: per-pixel segmentation, feature extraction, data preprocessing

• Reinforcement learning pipelines (data collection, training, deployment)

SYSTEM ARCHITECTURE & DESIGN

• Mediator + Observer patterns, single-source-of-truth state management

• Refactored 3500-line god classes to production-grade architecture (zero test regressions)

• Full-stack ownership: algorithm system design deployment CI/CD

• Production AI plugin frameworks (out-of-process, marshaling, callbacks)

• CMake build automation, self-contained deployments, Windows/Linux

DISTRIBUTED & REAL-TIME SYSTEMS

• 20+ years: C++ (C++11/14/17/20), Linux, Windows, embedded systems

• TCP/IP async I/O, message queues, shared memory IPC, DDS

• Multi-threaded debugging, race condition diagnosis, event-driven architecture

• ACE, CORBA, Boost (ASIO, Beast), native async patterns

• 985+ regression test suites, TDD-first development

EXPERIENCE

IMMERSIVEABZ – Senior C++ Engineer (Medical Imaging / 3D Visualization)

March 2026 - Present C++17, MSVC 2022, Qt 6.10, VTK 9.6, CMake, 985+ test suite

DICOMVIEW – Professional Windows DICOM Viewer with AI Segmentation

[FULL-STACK ARCHITECTURE & DEPLOYMENT]

• Sole engineer on production medical imaging viewer; end-to-end ownership

• Designed and refactored from monolithic architecture to Mediator + Observer pattern with single-source-of-truth

slice state; migrated without breaking any tests in 985-test regression suite

• Built AI segmentation plugin framework (out-of-process integration): TotalSegmentator payloads, ONNX pipeline

(planned); marshaling, IPC, result callbacks

• Owned complete build & deployment pipeline: CMake automation, vendored dependencies (Qt, VTK, GDCM, ITK),

self-contained Windows distribution with bundled runtimes, versioned releases

• Brought TDD and incident-driven engineering culture: captured 22 VTK + Qt OpenGL sharp edges as project rules

to prevent regressions (system-critical invariants)

[FEATURES & CAPABILITIES]

• MPR viewer end-to-end: synced 4-panel layouts, oblique reformats, volume rendering with transfer functions

• 2D/3D segmentation editor, measurement tools, crosshair & reference line sync across panels

• Multi-modal data fusion (medical imaging, annotations, segmentations)

SVAYA ROBOTICS – Software Engineer C++ [6-month contract]

January 2025 - April 2025 C++20, Boost ASIO & Beast, ROS2 Humble, DDS, Linux

ROBOT CONTROL & SENSOR INTEGRATION – PRODUCTION ROS2 MIDDLEWARE

[ROS2 PRODUCTION DELIVERY]

• Migrated Python robot control API to production C++ using C++20 coroutines and Boost ASIO

• Exposed robot's WebSocket interface as full-duplex ROS2 bridge for real-time telemetry and command routing

• Architected DDS middleware integration, message serialization, error recovery, real-time constraints

• Comfortable writing native async I/O: multi-threaded Boost ASIO event loops, concurrent client handling

[PROTOCOL BRIDGING & HARDWARE INTEGRATION]

• Built Modbus ROS2 bridge from scratch (multi-threaded over Boost ASIO): industrial protocol packet parsing,

state management, CRC validation, register mapping

• Built EthernetIP ROS2 integration: multi-sensor coordination, timing synchronization, failover logic

• Serial communication, hardware polling, real-time telemetry forwarding

• Deep experience integrating legacy industrial protocols with modern autonomous middleware

GAMING STUDIO, SYDNEY – Software Engineer C++ [3.5 years]

November 2020 - August 2024 C++11/14, VS2019, CUDA, HLSL, DirectX 11, OpenGL, Qt

GAMING SERVER ENGINE & GPU-ACCELERATED PERCEPTION PIPELINES

[GPU COMPUTE & MACHINE LEARNING]

• DirectX 11 & CUDA on NVIDIA hardware: GPU memory optimization, real-time constraints

• Implemented HLSL shaders for per-pixel segmentation, vehicle tracking, feature extraction; augmented render

buffers for downstream ML processing

• Built engine-side feature extraction feeding reinforcement learning data pipelines: segmentation IDs, vehicle

trajectory tracking, multi-modal feature vectors

• Profiling & optimization: GPU kernel tuning, batch processing, inference scheduling

• Python pybind11 API: memory-safe interop between game engine and ML frameworks

[TOOLS & ARCHITECTURE]

• In-house Qt UI tools for game developers: profiling, debugging, real-time monitoring

• Reinforcement learning experiments: vehicle navigation, navigation algorithm validation

BHP, PERTH – Big Data / Machine Learning Engineer [Contract]

March 2019 - May 2020 PyTorch, TensorFlow, Keras, OpenCV, scikit-learn, Spark, Kafka, OpenTSDB

COMPUTER VISION & SENSOR FUSION AT SCALE (50M+ records)

[COMPUTER VISION IN PRODUCTION]

• YOLO object detection on monitoring cameras; real-world mine site imagery

• UNet semantic segmentation for drivable area labelling: GPS-tagged training data, model evaluation, validation

• OpenCV image processing, data augmentation, annotation pipelines

• Model deployment and real-time inference

[SENSOR FUSION & STATE ESTIMATION]

• Time series forecasting: hybrid ARIMA + Kalman filter + RNN/LSTM/GRU for trailer battery prediction

• Multi-sensor data fusion, state machine implementation, uncertainty propagation

• 50M+ daily records; Spark SQL, distributed processing

[DATA ENGINEERING & INFRASTRUCTURE]

• End-to-end ingestion pipeline: NiFi stream processing, OpenTSDB time series storage, idempotent Kafka consumer

sinks (Java)

• MQTT-Kafka bridges, Web API integrations, Grafana dashboards

• Feature engineering, preprocessing, data quality pipelines

EVERTZ – Software Engineer C++

April 2025 - March 2026 C++17/20, wxWidgets, Vulkan, FFmpeg, Windows/Linux

OBD LINEAR PLAYOUT EDITOR – BROADCAST VIDEO

• Real-time editor for linear playout streams in StreamPro platform

• Text & 2D/3D animation pipeline; Vulkan graphics API; FFmpeg integration

IBM SECURITY DIVISION DEVELOPMENT LABS – Software Engineer [5.5 years]

August 2010 - March 2016 C++, Linux, AIX, network programming, virtualisation

NETWORK PROTOCOL ENGINEERING & TEST AUTOMATION

• IPFIX protocol encoder/decoder (UDP): shared memory real-time flow capture, QRadar SIEM integration

• C++ packet capture for IBM XGS Next Gen IPS appliance with CLI; host intrusion prevention

• Automated test infrastructure: QEMU virtualisation, Android/iOS emulators (Appium, Selenium WebDriver, REST

APIs) in Java & Python

• VMware ESX IBM AIX: ported virtualization code across platforms

PRE-2020 ROLES: System Architecture, Telecom, Finance

XALTED INFORMATION SYSTEMS, BANGALORE [Software Engineer C++, 2007-2010]

• Mediation & billing products: shared memory, message queues, IPC architectures

• Replaced Tuxedo with ACE + Boost + CORBA stack; cut licensing costs significantly

WEBODROME / ARISETECH [Lead C++ Engineer, 5 engineers, 2004-2007]

• Architected Automatic Call Distribution (ACD) system as CORBA services

• TCP/IP async I/O emulator on Linux in ACE: 10,000+ simultaneous client connections in testing

• IBM WebSphere MQ, MSMQ, DCOM-SOAP middleware

EMIND TECHNOLOGY / INFOSYS [UBS Investment Banking, 2003-2004]

• Securities trading & credit card backend systems

• C++ & Java on Solaris with ACE, CORBA, Oracle, IBM MQ Series

CYBAGE / BIGATES, PUNE & MUMBAI [C++ / Java, 1999-2003]

• 3GPP telecom: confidentiality & integrity algorithms, GPRS/3G packet generators

• CTI/Avaya call centre integration

• CORBA-based Network Management System (Catalant)

EDUCATION

Post Graduate Diploma in Data Science

IIIT Bangalore [2016-2017], CGPA 3.6 / 4.0

Diploma in Advanced Computing & Software Engineering

C-DAC, Business Computing-II [1998-1999], 89%

B.E. Industrial with Computers

R.K.N.E.C. Engineering College, Nagpur University [1991-1995]

Major in Operations Research

TRAINING & CERTIFICATIONS

Anthropic [2026]

Claude 101, Claude Code 101, Introduction to Claude Cowork, Claude Code in Action, AI Fluency Framework &

Foundations, Building with the Claude API

Machine Learning & Statistics

• Neural Networks for Machine Learning, University of Toronto (Coursera, 2017), 83%

• Bayesian Statistics: From Concept to Data Analysis, UC Santa Cruz (Coursera, 2017)

Systems & Middleware

• BEA Tuxedo training, Bangalore, 5 days [2008]

• CORBA C++ / Java workshop, iCMG India, 4 days [2000]

• UML workshop, CSI Mumbai, 2 days

TECHNICAL SKILLS

Languages: C++ (C++11/14/17/20), Python, C, Java, SQL, Bash

Robotics/Middleware: ROS2 Humble, DDS, Boost ASIO & Beast, ACE, CORBA, WebSocket

GPU & Graphics: CUDA, TensorRT, ONNX, OpenGL, Vulkan, DirectX 11, HLSL

ML & Computer Vision: TensorFlow, Keras, PyTorch, OpenCV, scikit-learn, YOLO, UNet, segmentation

Protocols & Networking: Modbus, EthernetIP, CAN, IPFIX, MQTT, Kafka, TCP/IP async I/O, serial comms

Real-Time Systems: Multi-threaded programming, TDD, event loops, 985+ regression test suites

Build & DevOps: CMake, Git, Docker, CI/CD automation, self-contained distributions

Platforms: Linux (20+ years), Windows (MSVC 2022), AIX, embedded, NVIDIA GPUs

PATENTS

• System and method to determine network access without user-provided identity credentials.

Filed 2015 [IBM]

• System and method to distribute and execute tests in large, agent-based, heterogeneous computer grids.

Filed 2013 [IBM]

• System and method for comprehensive metadata-driven testing of Web Services APIs and Application UIs.

Filed Q1 2016 [IBM]

References available on request.

Contact this candidate