www.linkedin.com/in/nimisha-agarwal-

***ab*** NIMISHA AGARWAL

(adsbxq@r.postjobfree.com)

Residency: Singapore PR

Nationality: Indian

+65 8375 5965

Seasoned scientist with experience in Product development, AWS, Video Analytics, Computer Vision, AI/ML, Digital Twins and 3D mapping. EDUCATIONAL QUALIFICATIONS

Nanyang Technological University PhD - Computer Science and Engineering [ Aug 2014 – Aug 2019] The LNM Institute of Information Technology B.Tech. – Electronics and Communication Engineering [ Jul 2010 – May 2014] PROFESSIONAL EXPERIENCE

Georgia State University, Atlanta, Georgia, USA Postdoctoral Research Scientist [Jul 2021 – Present]

Supervising and managing a team of 15+ developers, PhD students and interns.

Implementing and building data analytics solutions for hand-rehabilitation of paralytic patients by recommending them solutions for faster recovery.

Managing the development and deployment of ML pipelines for autonomous vehicles using EKF, SLAM and visual Odometry.

Delivered business pitch to prospective clients and hospitals to adopt the proposed solutions. Advanced Remanufacturing and Technology Centre

(ARTC), A*STAR

Development Scientist [ Aug 2019 – Present]

Implemented obstacle avoidance & precise localisation SLAM by using Intel Realsense from video feeds that will help in real-time 3D obstacle detection and avoidance for Auto-guided vehicles (AGVs) in ROS environment.

Recognized the objects using object recognition from video feeds to guide the AGVs for path detection.

Led a project on Proof-of-Concept Digital Twin implementation of the supply chain 4.0 aiming towards industry 4.0 in collaboration with Procter & Gamble

(P&G), Singapore to implement Digital Twin (3D simulation) solution; set Key performance indicators (KPIs) to achieve end-to-end visibility and real-time disruption management using supervised learning CNNs.

Deployed the Digital Twin app using Google Cloud.

Created a REST API call to send the data information about purchased items to either the Raw-material server or to the sub-assembly server to be stored in respective databases (InfluxDB, MySQL DB).

Designed a web-interface (UI/ UX) for user-interactive purchase order system that helped in prior suggestions based on a customer’s order history and to keep a track on purchase orders and display the purchased items on a webpage.

End-to-end development of AI based decision-aiding system for energy consumption of gearbox assembly.

Managed a team of 10+ researchers and full-time developers, providing guidance on technical and inter-personal aspects. Nanyang Technological University (NTU), Singapore PhD. Candidate [ Aug 2014 – Aug 2019]

Innovated new solutions and models on 3D RECONSTRUCTION: INVESTIGATION OF ACCURATE FRINGE ANALYSIS TECHNIQUES FOR FRINGE PROJECTION PROFILOMETRY (FPP)

Developed a structured light FPP system for 3D scanning using well-known calibration and image enhancement techniques.

First ever to incorporate GPU acceleration for fringe analysis for a high-speed 3D scan; achieved a real-time processing speed of 30 frames/sec.

Reconstructed 3D point cloud from 2D video frames at 30 frames/sec.

Mathematically analysed the issues associated with an FPP system such as shadowing around the curved areas etc.

Proposed novel computer vision techniques (both automatic and user-interactive) to eliminate these issues in MATLAB and C++. University of Trieste, Italy Research Associate [ May 2013 – Jul 2013]

Analysed H.264 codec to measure and reduce the blocking artefacts in video frames using DCT; monitored real-time video quality using C++. Indian Institute of Technology (IIT) Jodhpur, India Research Intern [ May 2012 – Jul 2012]

Proposed a novel edge preserving filter-based salt and pepper noise removal technique that helped remove more than 90% of the noise effectively. Indian Institute of Technology (IIT) Jodhpur, India Research Intern [ May 2011 – Jul 2011]

Proposed a novel least squares-based prediction scheme for image interpolation; achieved a reduction of upto 60-70% in computational efficiency. OTHER PROJECTS & EXPERIENCE

Engineered a face recognition and matching system (using TensorFlow) based on Caltech dataset that analysed the similarity quotient between two images to identify whether they are same or not.

Explored the possibility of AI in 3D measurements using neural networks.

Developed a user-interactive interface (using MeshView) to control objects using hand-gestures using Leap motion.

Embedded an image watermarking scheme to analyse the effect of different prediction schemes that can recover the original image later with 90% accuracy.

Awarded with Excellence in Teaching Assistantship by NTU for exemplary performance.

Awarded with NTU Research Scholarship from Aug 2014 to Aug 2018.

Served as a student leader (president) of Graduate Student Club at NTU for a year.

Organiser @ TedX NTU: Organised TedX events at NTU for 2 years consecutively.

Secured 2nd position in an intra-university hackathon competition at LNMIIT. SKILLS

Professional

Artificial Intelligence, Machine Learning, Deep Learning, 3D Computer Vision, 3D reconstruction and visualization, ROS Kinect, TensorFlow, Data Structures, Front End Development, Database Design Technical

Databases – SQL, NoSQL Adobe – Photoshop Microsoft – Word, Excel, PowerPoint Cloud Architecture –Azure, Google Cloud, AWS Programming – MATLAB, C/C++, Python, Node.js, HTML, CSS, JavaScript, React, CUDA, OpenGL

Languages English – Fluent Hindi – Native French – Beginner Relevant Courses Machine Learning Deep Learning Specialization Digital Image Processing Machine Vision PUBLICATIONS

S. Ri, N. Agarwal, Q. Wang & Q. Kemao

Comparative study of sampling Moire and windowed Fourier techniques for demodulation of a single fringe pattern Applied Optics [ 2018 ]

Q. Kemao, N. Agarwal, S. Ri & Q. Wang

Sampling Moire as a special windowed Fourier ridges algorithm in demodulation of carrier fringe pattern Optical Engineering [ 2018 ]

HOBBIES & INTERESTS

Badminton Swimming Music Reading Gym Travel