NIKHITHA BEKKANTI

Location: Michigan/ Remote E-mail: adx3hy@r.postjobfree.com

LinkedIn: https://www.linkedin.com/in/nikhitha-bekkanti/ Ph# 248-***-**** Education

• Master of Science in Computer and Information Science (Jan 2017-Dec 2018) University of Michigan, Dearborn GPA: 3.91/4

• Bachelor of Technology in Electrical and Electronics Engineering (Sep 2012-May 2016) Jawaharlal Nehru Technological University -Hyderabad, India Equivalent GPA: 3.9/4 Summary of Skills

• Modeling/Analysis Software: Python, Machine Learning (NLP, Computer Vision), C, Data structures, SQL, MATLAB

• Theoretical: Advanced Artificial Intelligence, Data Analytics in Software Engineering, Mobile Computing, Power Electronics, Hybrid electric Vehicles, Electrical machines, Electronic Devices and Circuits

• Inter personnel: Leadership skills, Team Player, Program Management, Project Scheduling and Project Management Work Experience

AI/ML Research Engineer-Intelligent In vehicle Experience -Ford Motor Company (Oct 2021 – June 2023)

• Developed an advanced machine learning model to predict the departure time of a driver based on their travel history. This model has potential applications in Battery Electric Vehicle (BEV) systems and consumer vehicle preconditioning.

• Developed a driver distraction prediction module to learn secondary activities such as drinking or texting, enabling the prediction of driver engagement duration and associated risk index. These modules have positioned the company at the forefront of delivering safe and intelligent features to the automotive industry.

• Led the integration of the karaoke feature into the Ford’s Intelligent In-vehicle driver alert technology (Invigorate platform). Developed an end-to-end process to seamlessly integrate karaoke with other driver engagement features including climate control, scent, seating position, and mood lighting.

• Collected and analyzed physiological data during the Invigorate study to assess driver alertness. Utilized these insights to propose an invigorating experience aimed at enhancing driver alertness and safety.

• Developed a machine learning model with 99% accuracy to predict driver engagement using facial landmark data Machine Learning Applications Research Engineer- Voice and Audio Technologies-Ford Motor Company (Jan 2019- Oct 2021)

• Built a machine learning model termed 'Domain Classifier' for Ford's in-house developed speech platform, CONVRS. This model is widely utilized across the company for various voice and audio applications.

• Conducted a comprehensive evaluation of two key suppliers for voice biometric authentication, leading to the establishment of dual-step authentication requirements for in-vehicle authentication and identification.

• Analyzed driver emotions in real-time using camera feeds and supplier SDKs. Proposed an approach based on valence/arousal quadrants to recommend in-cabin vehicle features tailored to the driver's emotional state.

• Developed a novel objective and subjective evaluation methodology to evaluate the performance of various vocal extraction softwares. This study was later used to drive the requirements for developing an in-house Ford vocal extraction software. Research Experience

Machine Learning Techniques Applied to a Real-World Test Data to Detect Human state (Jan 2018 - Apr 2018)

• Prediction of human state using semi-supervised learning of 'labeled train data' and 'unlabeled test data.'

• Developed a machine learning model with 80% accuracy by applying manifold regularization technique to 'LapSVM' algorithm. Design of a State Space Model of a Time-Series data using Markov Decision Process (MDP) (Jan 2018 - Apr 2018)

• Different MDP algorithms were studied, analyzed, and incorporated with the MDP toolbox (in MATLAB) to generate 'State Transition Probability matrix' from a given current state.

• Corresponding rewards are allocated for every state transition to improve the prediction efficiency. Electric Powertrain Development -Formula SAE (Society of Automotive Engineers) (Aug2017- Feb 2018)

• Worked on 2018 & 2019 Electric car under the University of Michigan, Dearborn FSAE team.

• Worked on the design of Electric power train and tuning of motor controller for better traction and performance.

• Worked on the data acquisition system of the motor using Arduino Mega with MATLAB programming. Patents/Awards/Honors

• Published and presented an article on “Evaluation of Voice Biometrics for Identification and Authentication” at SAE WCX-2021 conference https://www.sae.org/publications/technical-papers/content/202*-**-****/

• US Patent #US20200372908A1 Detecting and isolating competing speech for voice-controlled systems

• German Patent # DE102022107306A1 Systems and Methods to Improve the Sleep Experience in a Vehicle Cabin

• German Patent # DE102022130237A1 System and Method for Identifying Location of an occupant in a vehicle

• US Patent # US11644420B2 Occupant light exposure detection

• US Patent # US20220160144A1 Vehicle seating assembly (Variable heating)

• US Patent# Pending Intelligent and personalized karaoke for In-car applications

• US Patent# Pending Cross-vehicle entertainment using V2V and V2C2V technologies

• US Patent# Pending Analyze/predict driver emotional state to propose in-vehicle features

• Member of Student Union Board at UofM and was awarded ‘Best Outgoing Graduate’ in 2018

Contact this candidate