Jan 2024 – Dec 2024 · 11 mos · Cologne, North Rhine-Westphalia, Germany · Hybrid

• Developed AI models and evaluation strategies for autonomous railway operations, focusing on object detection, system assessment, and continuous model improvement.
• Designed and implemented a ROS2-based Data Recording System (DRS), handling architecture, sensor specs, and CI/CD integration for seamless data capture.
• Led research on AI model evaluation, authoring technical papers on system performance, FMEA strategies, and assessment metrics.
• Explored drone-based data collection feasibility, enhancing model training datasets with new data sources.
• Utilized AI prompting techniques (zero-shot, few-shot, chain of thought) with ChatGPT-4 and DALL·E to generate AI-driven visuals for project promotion.

Jun 2022 – Sep 2022 · 3 mos · Madrid, Community of Madrid, Spain · Remote

• Google Summer of Code 2022 participant with JdeRobot organization. My project focus on Optimization of Deep Learning models for autonomous driving. Achieved 50x improvement in inference time with the PyTorch optimized model from Tensorflow based baseline.
• Project webpage - https://theroboticsclub.github.io/gsoc2022-Nikhil_Paliwal/

May 2022 – Jul 2022 · 2 mos · Jena, Thuringia, Germany · Remote

• My work focused on optimizing deep learning models via various model compression techniques such as pruning, quantization, knowledge distillation, and so on. The task involves both academic research and code development to satisfy resource constraints and deploy models.

Nov 2021 – Sep 2023 · 1 yr 10 mos

• I am currently targeting the temporal generalization problem for the Twitter hashtag dataset. We are devising a contrastive pre-train network to improve performance over transformer based language models.
• My previous task was to devise a method for continual representation learning without the need for supervised signals. I created a self-supervision model for the problem of Ultra-fine entity typing. My responsibilities included doing additional research for the problem, implementing the code from papers, and performing further experiments and analyses.

Jul 2021 – Nov 2021 · 4 mos

• Part of the Interactive Machine Learning (https://iml.dfki.de/) research group working on multiple industrial research projects.
• I worked on creating a universal pre-train model for multiple medical tasks based on self-supervised learning algorithms.

Mar 2020 – Nov 2020 · 8 mos · Prague, Czechia · Remote

• The product is a health application to assist and train people to adopt proper sitting posture.
• I had the responsibility of developing an AI model to detect key points in the human body and use them to classify posture. Also, to conduct all related research and experiments.

Jun 2019 – Aug 2019 · 2 mos · Bangalore · On-site

• I was responsible for developing a smart video analytic system, implemented at various Reliance Petroleum Outlets. The work covers various use cases such as the identification of different vehicle types, human detection, vehicle counting and computation of the duration of individual events. Computer vision techniques such as CSRT tracker, and region filling, along with neural networks such as YOLOv3, were customized for the tasks.

Jan 2019 – Apr 2019 · 3 mos · Shenzhen, Guangdong, China · On-site

• Completed tasks on making intelligent robotics solutions for logistics industries. My work primarily focuses on the processing of point cloud and working with neural networks and computer vision algorithms to make robots aware of the complex environment with multiple RGBD and depth cameras.

May 2018 – Jul 2018 · 2 mos · New Delhi Area, India · On-site

• Worked on simulating a collision avoidance technique in gazebo using Collision Cone approach for multi-UAV’s control with ROS under the guidance of Dr P.B. Sujit.

Dec 2017 – Dec 2017 · 0 mo · Taiwan

• Implemented and analyzed various Artificial Intelligence techniques for enhancing the experience of different multimedia means and devices.

Aug 2017 – Nov 2017 · 3 mos

• Guided 2nd year undergraduate students in implementing various algorithms used in Advanced Programming during laboratory sessions.

Jun 2017 – Jul 2017 · 1 mo

• The project idea was to automate the whole process of vehicle passing that can later be used on various places like traffic signals, tolls, schools, colleges etc. Worked on different algorithms like background subtraction, frame subtraction, haar cascades, HOG feature and different OCR techniques to implement a working and efficient system.

Jan 2017 – Apr 2017 · 3 mos

• Guided 1st year undergraduate students in implementing various algorithms,taught in theory classes using C language during laboratory sessions.

Aug 2016 – Nov 2016 · 3 mos

• Conducted various laboratory sessions, test and evaluation to help students grasp concepts of computer programming

Jun 2016 – Jul 2016 · 1 mo

• This program gives huge exposure and practical experience of various tools and design techniques used in field of VLSI. Some mentionable tools like Vivado used for VHDL circuit design, FPGA's like Zybo, Spartan 3 , Embedded system boards like Intel Galileo, Beagle Bone, Embedded Artist, Picoblaze etc. were used to design many practical solutions.

Jul 2017 – May 2018 · 10 mos

• This club gives technical training and guidance to enthusiastic students in the field of Robotics, Computer Vision, Electronics and Embedded System. Being the coordinator of the club, I have given technical lectures, laboratory sessions in different technical fields. We have also conducted various technical competitions in and outside the institutes to provide a technical platform for interested students. I have led many students teams for different technical competitions. This club is a hub of technical enthusiasts who are really willing to learn and contribute to society. I have taken care of every aspect of this club with all my will and dedication.

Jan 2016 – Jan 2016 · 0 mo · Jaipur Area, India

• Workshop Organizer,my work include, from maintaining database of all companies to finalising the workshop and onspot managing the workshops also. We finally managed to organize some fantastics workshop on topics like Computer Vision, Web Development, Astronomy etc.
• Robotics event organizer,( Drone Obstruction). This was a completely new event in Plinth, we managed to bring participation from all over India which include teams from colleges like Nit Jalandhar, GIT gandhinagar, VIT vellore etc. The event was huge success.

Jan 2015 – Jan 2016 · 1 yr

• Worked with Firebird V Atmega2560 and Spark V Atmega16A Robotic Research platform. Creating a smart solution under certain constraints of competition using different sensor like Sharp, IR distance,Color sensors etc.