Cambridge, Massachusetts, United States
Developing the next generation of advanced driver assistance and autonomy features for software-defined vehicles, as a member of the Motion Planning team.
Led the development of advanced autonomy features, including high speed lane-change planning, enabling smoother, safer L2+ driving.
Filed multiple patents on scalable autonomous lane change planning systems.
Selected for cross-regional talent exchange in Japan; onboarded local planner teams to core autonomy workflows, unified lane change system understanding across US/JP teams, and accelerated bringup of new production platform through hands-on triage and architectural alignment.
Optimized runtime performance of the autonomy stack, reducing latency in trajectory planning by over 40%, ensuring real-time responsiveness.
Integrated the core Trajectory Optimizer library into the L2+ driving app, which involved porting thousands of external Bazel targets.
Drove cross-functional integration of perception, planning, and control modules for complex driving scenarios.
Enhanced debugging, testing, and visualization tools, accelerating the development cycle.
Promoted compliance with safe coding guidelines as a Platform Reviewer; mentored other engineers, fostering growth and contributing to the team’s technical excellence and knowledge-sharing culture.
Waltham, Massachusetts, United States
Responsible for developing, integrating, testing, deploying, and maintaining the
autonomy stack of Vecna’s suite of mobile heavy material handling warehouse robots.
As the module owner for navigation, developed best-in-class path planning and
obstacle avoidance features which significantly improved speed and robustness in
tight, dynamic spaces, thereby maximizing uptime and throughput. Achieved a 135% increase in overall performance composite scores and a 130% improvement in autonomous recovery success rate.
As the software technical lead, led a multidisciplinary team to design, develop, and integrate autonomy software for Vecna Robotics’ Co-Bot Pallet Jack (CPJ) – an industry first Autonomous Mobile Robot (AMR) that combines small size, affordability, and advanced shared autonomy capabilities to disrupt traditional pallet transportation workflows.
Led a team in developing autonomy software for a novel mobile manipulation robot designed for robotic shelf picking in warehouse environments. The Tote Retrieval & Storage (TRS) Robot was engineered to autonomously pick, place, store, and transport packages and totes across a warehouse, unlocking a previously untapped segment in the market. Won first place at the DHL & Dell Robotics Innovation Challenge, and demonstrated a working prototype at MODEX 2018 (the global supply chain trade show), thereby securing funding from interested customers to build a deployable product.
As the C++ domain owner, discussed and documented best practices, encouraged
their use via code reviews, and emphasized removal of tech debt.
Brought key improvements to pallet docking behaviors which increased pallet
handling reliability and throughput in long, densely packed lanes.
2017 — 2018
Cambridge, Massachusetts, United States
Played a pivotal role in the development of a dual-arm mobile manipulator designed for safe and remote Explosive Ordnance Disposal (EOD). This robot combined autonomous and teleoperated capabilities to enable effective operations in high-risk environments.
3D Teleoperation Interface: Designed and implemented a user-friendly 3D teleoperation interface using C++, ROS, RViz, and MoveIt libraries, significantly reducing the cognitive load on operators.
Advanced Grasp Detection: Integrated a cutting-edge grasp pose detection library that used convolutional neural networks (CNN) on point cloud data, enabling autonomous picking of arbitrary and novel objects in cluttered environments without CAD models.
Autonomous Behaviors: Developed high-level autonomous behaviors to simplify teleoperation, replacing earlier reliance on direct joint control, which was imprecise and operator-intensive.
Enhanced Situational Awareness: Improved operator awareness by developing a UI that visualized depth information, providing a comprehensive view of the robot's surroundings.
Delivered successful project milestones and customer demonstrations, showcasing reliability and potential real-world applications, and securing additional funding from the customer to advance the platform's capabilities and readiness.
2016 — 2016
San Francisco Bay Area
Performed exploratory R&D and rapid end-to-end prototyping for a new kind of headphones with ungrounded force actuators to provide instinctive, non-visual and non-auditory feedback to the wearer. The system can be used to guide the wearer in a subtly (e.g., pedestrian navigation) or to focus the wearer’s attention in a specific direction (e.g., personal warning system) without interfering with the already heavily stimulated sensory pathways of sight and sound. Planned and created fully working prototypes, including hardware and software, showcasing the user interaction described above. This included:
Investigating and sourcing appropriate hardware technology — microcontrollers, peripherals, actuators, sensors, power sources, and structural components — to test different concepts.
Implementing precise mechanical systems for center-of-gravity-shifting and other ungrounded force actuation techniques, from CAD to 3D printing to mechanism assembly.
Designing and miniaturizing electronics to control mechanical systems and synchronize them with other software components.
Developing accompanying embedded software as well as Android apps to allow for basic interaction with the system and to demonstrate its usability.
Darmstadt, Hesse, Germany
Undergraduate thesis on lower-limb wearable robotics at the Locomotion Lab, Institute of Sport Science.
Thesis title: Gait Analysis and Control Design for Stair Ambulation with Lower-limb Powered Prostheses
Advisors: Dr. Andre Seyfarth, Dr. Martin Grimmer
Investigated lower limb joints and segments during level walking, stair ascent, and stair descent to determine their biomechanics for use in the design of new and robust powered prosthetic systems which can efficiently negotiate stairs in addition to walking on flat ground.
Performed a motion-capture experiment with an instrumented staircase setup on a healthy subject.
This study contributed two new control insights for gait intent and gait percent detection in wearable lower-extremity robots which are not available in the existing literature.
Significant savings in motor peak power and energy requirements across all gaits were obtained by optimizing the spring stiffness of the Series Elastic Actuator (SEA) based powered ankle prosthesis, which allows motor and battery dimensions to be kept to a minimum – an essential requirement of wearable robots.
The results showed that current powered prostheses can extend their capability beyond level walking to successfully negotiate stairs, with only minor modifications and sensor additions to their existing systems.
Education
Northwestern University
Master of Science (MS)
Birla Institute of Technology and Science, Pilani
Bachelor of Engineering (B.E. Hons.)
Ahlcon International School