Kshitij Srivastava
Education
Indraprastha Institute of Information Technology - Delhi, IN
July 2018 - July 2020
M.Tech (CSE), CGPA 8.33
Thesis: Applications of Open Transit Data
Awards:
- Appreciation for Masters Thesis
- Dean’s R&D Award - 2019
- AICTE PG Scholarship - 2018-20
Activities:
- Student Mentor for 2019-20 - mentored a group of 10 Masters students.
- TA for Discrete Mathematics, Data Warehousing, Introduction to 3D design, Privacy and Security in Online Social Media
Dr APJ Abdul Kalam Technical University - Lucknow, IN
August 2013 - June 2017
B.Tech (CSE), 74.44%
Activities:
- Placement coordinator for 2015-16
- Head Coordinator for Renaissance committee 2015-16. Responsible for organizing the runway event and managing sponsorships
Experience
Research Assistant, Mobility and Optimisation Lab - IIIT Delhi, IN
Nov 2020 - Current
- Working on GTFS data creation (static and real-time) for the two transit agencies in Delhi (DIMTS and DTC)
- Responsible for enabling the integration of this data with third party vendors
- Part of the ticketing team conducting e-ticketing trials across Delhi
Consultant, Mobility and Optimisation Lab - IIIT Delhi, IN
May 2020 - Oct 2020
- Worked on a project to lay down policy guidelines for creation and dissemination of Open Data for Public Transport - Chennai for the World Bank
Graduate Student Researcher, Mobility and Optimisation Lab - IIIT Delhi, IN
Jan 2019 - July 2020
- Performed a study to evaluate the effect of the odd even scheme on traffic congestion in Delhi
- Analyzed the impact of bus breakdown on traffic congestion
- Developed a system to detect bus stop points in real-time using spatio-temporal clustering
Intern, Chartr - IIIT Delhi, IN
May 2019 - July 2019
- Worked on developing a multi-modal trip planning algorithm for Delhi based on static data
- Integrated this algorithm with a Django API module to be served to an android app
Intern, Tata Consultancy Services - Lucknow, IN
June 2016 - Sept 2016
- Developed an HTML UI Generator tool
- The tool helped non-programmers to generate desired HTML pages
Publications
Srivastava, K., Nair, S., Malik A. & Biyani, P. (Submitted). Bus Breakdown Detection Using Real-Time Transit Data. In 2021 IEEE ITSC.
Girsa, R., Srivastava, K., Jain, A., & Biyani, P. (2021, January). Demo Abstract: Contactless E-ticketing in Public Transport Buses. In 2021 COMSNETS. [Link]
Srivastava, K., & Biyani, P. (2020, August). Applications of open transit data (Master’s thesis, IIIT-Delhi). [Link]
Projects
Public Information System : Public Transport Data, GTFS [Link]
- Developed a plug-n-play public information system for commuters
- The information can be displayed on any display using a web browser
- Made the backend and required APIs for the application to work
Effects of Odd-Even Policy on traffic speeds in Delhi : Policy Evaluation, Open Transit Data, GTFS [Link]
- Studied the impact of the odd-even road rationing policy on traffic conditions in Delhi
- The study was performed using open transit data
- Observed a consistent decrement of 15-20% in travel times owing to the policy
- Work covered by major national newspapers)
Assessing the impact of bus breakdown on fuel costs : Machine Learning, Clustering
- Collected data for bus breakdowns from bus depot
- Using the open GTFS data of buses, categorized breakdown cases
- Used spatio-temporal DB-SCAN to detect these breakdowns in real-time
- Employed a fuel consumption model to estimate the cost of the extra fuel consumed
Projects
Voice Conversion : VCC'16, Deep Learning
- Used CycleGAN-VC2 and HMM for parallel and non-parallel voice conversion
- Trained the models on VCC’16 and CMU Arctic dataset
- Fine tuned models for performing the task on Malayalam Speech dataset from OpenSLR
Economy mapping using Satellite images : Multi-task learning, Deep Learning
- Classified the economy of a region using satellite images, combined over sub-districts in Uttar Pradesh, India
- Used a three headed multi-task learning model to predict income levels of a region from satellite images by detecting developmental features such as water, lighting source, etc
- Based on the predicted income levels, classified the region into economically sound or poor using a different neural network
Detection of Road Violations via Mobile Phone : Android Development, Data Collection
- Objective was to detect traffic violations at road cross sections using a mobile phone’s camera
- Worked on the design and functionality of the android app
- Collected extensive ground data from several traffic crossings
Predicting Dengue Spread : Machine Learning, Regression, Boosting
- Predicted the number of dengue cases in two cities based on given environmental conditions
- Applied linear regression and boosting techniques such as XGBoost
Genetic CNN : Genetic Algorithms, Convolutional Neural Network
- Designed two binary encodings for representing a Convolutional Neural Network
- Applied Genetic Algorithm to find the best CNN architecture for image classification on MNIST dataset
- Further tested the best CNN architecture trained on MNIST dataset on CIFAR10 dataset and achieved an accuracy of 98.39 %
Violence against Women - Prevention : Data Collection, Analysis, Visualization and Association Pattern Mining
- Collected data from various government websites for crimes against women and their probable factors
- Segregated Indian states based on ‘Crimes Against Women’ to ‘Total Crimes’ ratio and classified them into safe, moderately safe or unsafe according to predefined thresholds
- Generated transactions for some states including crimes and the probable factors behind them
- Applied Apriori algorithm to generate association rules from these generated transactions
Skills
Programming Languages : Python, Shell scripting
Frameworks : Django, Flask, Pandas
Tools : Jupyter, Git
Languages
Hindi : Native
English : Fluent, TOEFL - 113/120