Image Captioning with CNN-RRN Transformer Encoder-Decoder
→
Summary
Developed an image captioning system leveraging a CNN-RRN Transformer Encoder-Decoder architecture for advanced image understanding and natural language generation.
About
A highly analytical and results-driven Computer Vision and Robotics Engineer with a Master's in Autonomous Systems, specializing in real-time sensor fusion, deep learning inference optimization, and embedded hardware design. Proven ability to develop and validate complex perception and control systems for autonomous vehicles and robotics, leveraging expertise in CNNs, Kalman Filtering, and CI/CD pipelines to deliver high-accuracy, real-time solutions.
Work
SUPERFLASH
|Computer Vision Engineer
N/A, N/A, Finland
→
Summary
Developed and optimized advanced computer vision systems for autonomous vehicles, focusing on real-time performance and high accuracy in localization and perception.
Highlights
Developed a cutting-edge Visual Odometry (VO) system, integrating CNN-depth estimation with geometric methods to enhance spatial understanding for autonomous navigation.
Optimized Deep Learning (DL) inference pipelines on GPU, achieving real-time performance and efficiency through model quantization and advanced memory management techniques.
Reduced Point Cloud Library (PCL) outliers by implementing an adaptive filtering algorithm, significantly improving Autonomous Vehicle (AV) localization accuracy and reliability.
Maintained high position accuracy across diverse autonomous navigation paths, contributing to robust and safe system performance in dynamic environments.
Validated the perception system against rigorous industry-standard datasets (KITTI, TUM, EuRoC), ensuring compliance, robustness, and high-fidelity performance.
Implemented comprehensive CI/CD pipelines utilizing Docker containers and Gitlab, integrated with Jira/Agile project management for streamlined development and deployment workflows.
VECTOR MACHINES
|Robotics Engineer - Sensor Fusion Intern
N/A, N/A, Finland
→
Summary
Engineered and validated a proof-of-concept for hybrid localization, integrating multi-sensor data for robust navigation in GPS-denied environments.
Highlights
Developed a Proof-of-Concept (POC) for hybrid localization, enabling robust software deployment and reliable navigation in GPS-denied environments.
Implemented real-time sensor fusion systems, integrating LiDAR, IMU, and GNSS RTK data with Kalman Filtering for enhanced state estimation and real-time data analysis.
Adapted and optimized open-source codebases through extensive simulation, bench, and field testing, ensuring system robustness and high-performance operation.
Maintained high position accuracy across diverse autonomous navigation paths, critical for precise robotic control and operational reliability.
Established and managed CI/CD pipelines utilizing Docker containers and Gitlab, integrated with Jira/Agile methodologies for efficient project lifecycle management.
Collaborated on strategic research and commercialization initiatives, contributing to the successful development and market readiness of innovative products.
AALTO UNIVERSITY
|Research Assistant
Espoo, Uusimaa, Finland
→
Summary
Conducted research and development for Maritime Autonomous Surface Ships (MASS), focusing on collision avoidance and real-time control systems.
Highlights
Developed advanced collision avoidance software specifically for Maritime Autonomous Surface Ships (MASS), enhancing safety protocols and operational efficiency in complex marine environments.
Debugged and seamlessly integrated complex hardware-software systems to achieve precise real-time vehicle control for autonomous marine applications.
Executed rigorous bench and field testing in controlled ice tank environments under harsh maritime conditions, critically validating system performance and resilience.
Optimized Kalman Filter algorithms, significantly reducing sensor noise and improving state estimation accuracy by over 15% for enhanced navigation.
Contributed to a peer-reviewed publication and comprehensive system documentation, effectively disseminating research findings and ensuring knowledge transfer.
TRITON ELECTRIC VEHICLE INDIA
|Embedded Hardware Engineer
N/A, N/A, India
→
Summary
Designed, integrated, and tested critical embedded hardware systems for electric semi-trucks and bikes, ensuring compliance with industry standards.
Highlights
Designed and integrated critical power electronics, including DC-DC converters, Inverters, BMS, MCUs, and PDNs, for electric semi-trucks and bikes, enhancing vehicle performance.
Managed the complete hardware design cycle, encompassing requirements analysis, rapid prototyping, component selection, schematic design, and rigorous validation for EMI/EMC compliance.
Performed extensive bench and field testing and debugging using advanced tools like oscilloscopes, multimeters, SMPS, and logic analyzers, ensuring robust system functionality.
Implemented CI/CD pipelines, project management within Jira/Agile frameworks, and version control, ensuring efficient development and EMI/EMC compliance.
Collaborated effectively with Mechanical, Software, and Stakeholder teams, driving product development and comprehensive documentation efforts.
Education
EIT DIGITAL
N/A, N/A, Finland
→
Master of Science
Autonomous Systems
Courses
Computer Vision
Machine Learning
Deep Learning
Deep Network Development
Image and Signal Processing
Languages
English
Skills
Programming Languages
C, C++, Python.
Robotics & Autonomous Systems
ROS2 Humble, Sensor Fusion, Visual SLAM, Autonomous Navigation, Real-time Systems, Embedded Systems, Hardware-Software Integration, CI/CD, Jira, Agile, Docker, Gitlab, Kalman Filtering, Robot Localization.
Computer Vision & Deep Learning
PyTorch, OpenCV, CNN-Depth Estimation, Deep Learning Inference Optimization, Model Quantization, Object Detection, Image Captioning, Hough Transform, Image and Signal Processing.
Hardware & Electronics
GPUs, Altium Designer, LTSpice, Raspberry Pi, ESP32, Jetson, Arduino, Circuit Design, PCB Design, DC-DC Converters, Inverters, BMS, MCU, PDN, EMI/EMC Compliance, Bench Testing, Field Testing, Debugging Tools.
Software Development & Operations
Model Deployment, Performance Validation, Quality Metrics, Version Control, Continuous Integration, Continuous Deployment.
Projects
Object Detection with Vision Model for Autonomous Mobile Robot
→
Summary
Engineered an object detection system utilizing a vision model specifically tailored for deployment on an autonomous mobile robot, enhancing its environmental perception.
Visual SLAM with Hough Transform implementation from scratch
→
Summary
Implemented a Visual Simultaneous Localization and Mapping (SLAM) system from scratch, incorporating Hough Transform for robust feature extraction and environmental mapping.