Academic and Independent Projects

All my projects can be found at Github. Below mentioned are some of the important ones.

Optimizing Zero-Check Protocol

• Research Project @ New York University (Jan. 2025 - Present)
• Working along side researchers at Brooklyn Application, Architecture and Hardware Lab (BAAHL) in implementing the zero check protocol using Arkworks, to check if a univariate polynomial evaluates to 0 on a given set of field elements optimized for hardware accelerators
• Optimizing and benchmarking the performance against sum-check based zero-check protocol on composition of multi-linear functions introduced in [CBBZ’23] ’HyperPlonk: Plonk with Linear-Time Prover and High-Degree Custom Gates’
• Skills: Arkworks, ZK-SNARKs, Rust[Code]

Doubly-Efficient Interactive Proofs

• Research Project @ New York University (Sept. 2024 - Present)
• Studying the Doubly-Efficient Interactive Proof systems proposed earlier in [GRK’08], [RRR’15], and [GR’16], and researching on existence of classes beyond N C and SC for which doubly-efficient proof systems exist
• Studying the requirements to refute the Non-deterministic Strong Exponential Time Hypothesis (NSETH) using the assumption that k-OV ∈ DE-IP and de-randomization assumptions
• Studying the barrier in achieving efficient non-interactive proofs beyond a certain communication threshold
• Skills: Interactive Proofs, Computational Complexity Theory

Quantum-Secure Accumulation Schemes

• Research Project @ New York University (Sept. 2024 - Dec. 2024)
• Implemented the accumulation scheme without homomorphic commitments as suggested in [BMNW’24] ’Accumulation without Homomorphism’ using quantum-secure Merkle trees and error correcting codes using Arkworks
• Reviewing relevant literature to develop and optimize the Accumulation scheme by accumulating proofs in a tree-like folding scheme
• Implemented and optimized [BMNW’24b] ’ARC: Accumulation for Reed-Solomon Codes’ using Arkworks
• Skills: Accumulation Schemes, Arkworks, Rust [Code]

Collector Coin by NYU, DLM Group

• Academic Project @ New York University (May 2024 - Aug. 2024)
• Worked as a part of a 2-member team in creating a market place for vintage car restoration and trading, backed by blockchain
• Done as a part of the 'Info Tech Projects' course
• Implemented features for restorers and buyers to bid for projects and cars through smart contracts
• Enabled investors to invest on restoration projects using a new ERC20 token deployed on Ethereum
• Skills: Java, Spring Boot, Blockchain, Solidity, ReactJS

Zero Cash Sapling Protocol using Arkworks

• Academic Project @ New York University (Jan. 2024 - May 2024)
• Worked as a part of a 2-member team in implementing the ZCash Sapling Protocol in Rust using the Arkworks library
• Done as a part of the 'Cryptography for Blockchain' course
• Skills: Rust, Cryptography, Blockchain, Zero-Knowledge Proofs, Arkworks [Code]

Defense Against Evasion Attacks in Machine Learning Models

• Research Project @ PSG Tech (Jan. 2023 - May. 2023)
• Simulated evasion attacks on image classification model, under various assumptions on the amount of prior knowledge known to the attackers, about the system
• Analysed the effectiveness and efficiency of the evasion attacks on the classifier and its impact on the performance of the model in terms of the confidence score, precision, robustness and F1-score
• Proposed a defense mechanism involving retraining of the model repeatedly, adjusting the weights, to classify even the adversarial examples of the evasion attack, correctly
• Skills: Machine Learning, Security, Python, PyTorch [Code]

Privacy Preserved and Secure Federated Learning

• Research Project @ PSG Tech (Aug. 2022 - Nov. 2022)
• Developed a Federated Learning model to overcome the security limitations of the standard machine learning models, in recognizing patients suffering from dyslexia by processing details such as handwriting
• Incorporated privacy-preserving mechanisms such as Differential Privacy, Homomorphic Encryption, and Secure MultiParty Computation into the federated learning model
• Analyzed the accuracy achieved for the developed model against different levels of privacy, through fine-tuning the amount of Laplacian and Gaussian noise introduced to the gradients
• Skills: Federated Learning, Machine Learning, Cryptography, Python, PyTorch, TenSEAL [Code]

Food Recipe Recommendation System

• Academic Project @ PSG Tech (Aug. 2022 - Nov. 2022)
• Worked in a 3-membered team as a part of 'Social and Economic Network Analysis' course
• Studied and implemented a recommendation engine based on the 'Collaborative Filtering' algorithm to recommend recipes to users
• Designed a GNN to map the macro-nutrition values to the ingredients and suggest healthy alternate ingredients for recipes as per the FSA recommendations
• Skills: Knowledge Graphs, Social Network Analysis, Machine Learning, Tensorflow, Scikit-Learn, Python, Gephi [Code]

Unified Parsing Script using Machine Learning

• Academic Project @ PSG Tech (Jan. 2022 - May 2022)
• Headed a team of 6 to overcome the issues associated with the primitive web scrapers such as being site-dependent and highly rigid to changes in the HTML template
• Used three independent machine learning models that recognize patterns in the position of the desired information in the website, it's dimensions and the HTML tag enclosing it respectively
• By embedding the three models together, developed a site independent machine learning system, that extracts key information from different product pages from a variety of e-retailer sites
• Achieved an accuracy of 96% in extracting the key information from previously seen sites and an accuracy of 82% in extracting information from previously unseen web sites
• Skills: Machine Learning, Web Scrapping, Selenium, PyTorch, Python [Code]

Automatic Hand Sanitizer Dispenser

• Academic Project @ PSG Tech (Apr. 2022 - May 2022)
• Built a prototype of an automatic hand sanitizer dispenser as a part of 'Embedded Systems' coursework
• Primarily made use of ultrasonic sensors, relay module,s and pumps to build the real-time system
• Skills: Arduino, C++ [Code]

Vaccine Drive Registration System

• Academic Project @ PSG Tech (Jul. 2021 - Nov. 2021)
• Worked as a part of a 6-membered team in developing an efficient inoculation appointment system
• Project was developed as a part of 'Software Package Development' course
• Developed using Java, Java Swing, MySQL & Python and involved a machine learning model to predict the crowd at hospitals in order to avoid over-crowding
• Skills: Machine Learning, Python, Java, Java Swing, MySQL [Code]

Bulls and Cows Game via LAN

• Academic Project @ PSG Tech (Oct. 2021 - Nov. 2021)
• Worked as a part of a 5-membered team in developing a digitalized version of the old bulls and cows math puzzle
• Designed the game to allow multi-player support by connecting users through local ports using a TCP protocol
• Skills: Computer Networks, TCP, UDP, C [Code]

SportsQuip - Web Site

• Hackathon Project @ PSG Tech (Jun. 2021 - Jun. 2021)
• Developed as a part of the Front-End Web Development Hackathon Hack2k21 hosted by E-Cell of IIIT, Pune
• Developed a one-stop sports equipment purchase site in accordance with the topic provided and secured 2nd place in the 48-hour International hackathon
• Skills: Front-end Development, HTML, CSS, JS, Bootstrap [Code]

Bus Ticket Reservation System

• Academic Project @ PSG Tech (Jun. 2021 - Jun. 2021)
• Worked as a part of a 6-membered team in a developing a bus ticket reservation system
• A DBMS based mini-project developed using Java, MySQL & Java Swing
• Skills: Database Management Systems, Java, Java Swing, MySQL [Code]

Performance Enhancement using Multi-Threading

• Academic Project @ PSG Tech (Jun. 2021 - Jun. 2021)
• Analyzed the performance gain achieved through multithreading as part of 'Operating Systems' coursework
• Studied the relationship between the execution time and level of parallelism and function parameters for Quick Sort and Merge Sort applications
• Skills: Operating Systems, Multi-Threading, C [Code]