Hello, I am

Nalin Bendapudi

I am a

About

About Me

  • Name: Nalin Bendapudi
  • Location: Ann Arbor, MI, USA
  • Current Role: Autonomous Vehicle Mapping and Localization Engineer, Ford Motor Company
  • Interests: SLAM, Sensor Fusion, Perception, State Estimation, Motion Planning

Education

2019-Present

Master of Science in Robotics

University of Michigan, Ann Arbor

Coursework: Robotics Systems Lab, Self-Driving Cars, Mobile Robotics (SLAM), Robot Operating Systems, Deep Learning for Computer Vision, Linear Feedback Control, Digital Control, Robot Modelling and Control, Hybrid Control, Motion Planning

2013-2017

Bachelor of Technology in Mechanical Engineering

Indian Institute of Technology, Delhi

Coursework: Machine Learning, Data Structures, Analysis and Design of Algorithms, Embedded Systems, Robotics, Dynamics of Multi-body Systems, Control Systems

Experience

June 2021 - Present

Ford Motor Company, Ann Arbor

Autonomous Vehicle Mapping and Localization Engineer
  • Implemented third-party-map-based visual localization algorithm using pose-graph optimization technique to integrate IMU, radar,and GNSS measurements with a distance-transform image based map-matching constraint
  • Developed a real-time application to project map information on camera image to monitor calibration and localization accuracy, and aid the object-detection network by providing an initial bounding-box estimate of traffic signs
  • Implemented and bench-marked a lane assignment algorithm to determine the accurate lane-id of ego and traffic vehicles by fusing nearest lane candidates from map and the planned-path lanes in a Hidden Markov Model approach
  • Designed a grid-map library to load semantic map layers in real-time and provide functions to query map information
  • Surveyed various Visual-Laser-Inertial fusion algorithms and presented a new design for vehicle's localization system
  • Assisted in various stages of HD-Map creation: collected data, built pose-graph, constructed lane-level RoadRunner map
  • Configured and calibrated GNSS and IMU sensors; modified their ROS drivers according to vehicle's software architecture
  • Developed metrics and visualization tools to test and analyze performance of individual sensors and the whole sub-system
January 2021 - April 2021

University of Michigan, Ann Arbor

Graduate Student Resarch Assistant
  • Developed novel simple policies and integrated them in the Decentralized Multi-Policy Decision Making (D-MPDM) framework to plan efficient paths for a robot cluster executing an adversary search task under uncertain communications
  • Enhanced the quality of existing D-MPDM code-base by improving modularity and refining the documentation
August 2017 - May 2019

Robert Bosch Engineering and Business Solutions, Bengaluru

Senior Engineer, Automation Team
  • Analyzed trend and seasonality of time series data, and developed applications for anomaly detection and forecasting of automobile metrics. The applications were deployed on Mercedes-Benz servers and evaluated to have an accuracy of 92%
  • Utilized deep-NLP models like Google's BERT on unstructured error logs to develop a recommendation engine to identify potential high-impact issues
  • Applied unsupervised learning to find associations among data which helped in reduction of resolution and root-cause analysis time by 40%
July 2016 - July 2017

Indian Institute of Technology, Delhi

Undergraduate Teaching Assistant
  • Mentored three batches of 20+ sophomores in interpreting observations and reporting results in Materials Science Laboratory
  • Conducted doubt-clearing sessions to review tutorial problems and helped students visualize concepts using experiments
  • Graded assignments, home-works, and quizzes and proctored and organised both practical and theory examinations
May 2016 - July 2016

TBK Co. Ltd. (Japanese OEM), Tokyo

Summer Intern
  • Reverse-engineered an S-Cam air brake to calculate and correlate net brake torque, shoe factor and maximum pressure
  • Disassembled and executed a Fault Tree Analysis for cavitation, rust, cracks, dents and fretting in E694 Eicher Pump
  • Researched critical parameters, Pokayoke mechanisms & automation principles in brake and pump manufacturing

Skills

C++

90%

Python

85%

MATLAB

75%

ROS

80%

Projects

Projects

Blog

Blog

Is belief a scientific concept or a religious one?

January 4, 2021

Belief is quite a scientific concept. But the battle between science and religion is still justified.
(p.c. : Probabilistic Robotics, S. Thrun)

Perfect Memory stands in between Computers and Artificial General Intelligence

July 11, 2020

Computers still can't think. It's not because of how they process data but how they store it.

What is it?

July 6, 2017

If you find out what it is, try to find something even more interesting (Hint: a pattern) in the poem. If you have no idea what it is, this pattern will give it away!

Mathematical Motivation for Belief in God

Mar 23, 2017

Is it mathematically justified to believe in God although we don't have any scientific proof of God's existence?

High On Entropy

June 23, 2014

The poem will obviously seem quite senseless and completely random (hence the title). The poem can be better understood by a listener than the reader

Contact

Contact Me