Sanjeev Vijayakumar Email & Phone Number

Sunnyvale, Ca, Us

Las Vegas, Nevada, United States

Rochester, New York, United States

Amsterdam, Netherland

In collaboration with the University of Amsterdam, me and my partner addressed plastic waste and environmental sustainability through machine learning for predicting oil yields in plastic hydrothermal liquefaction. Recycling plastics holds promise, but the variability in trash-derived feedstock complicates oil yield predictions. I created a comprehensive dataset, including key parameters like temperature, pressure, heat rate, and plastic type.After evaluating ten research papers, XGBoost was chosen as the optimal model, achieving a Mean Average Error of 23.45%. One-hot encoding of the categorical variables assisted in optimization. I fine-tuned the model based on an advanced literature review, highlighting the potential of catalytic pyrolysis and optimal temperatures like 425°C for PS pyrolysis.The project culminated in an 80-page thesis, earning commendable grades, showcasing my commitment to sustainability and interdisciplinary research, and demonstrating my expertise in ML.

Toulouse, Occitanie, France

My two-month summer internship in France was a transformative journey in the field of artificial intelligence. I focused on enhancing Dense Retrieval Models, under legal information retrieval. The lab's 'Lawbot' division aimed to efficiently extract vital information from extensive legal documents, addressing a crucial need in the legal sector.My role specialized in optimizing Dense retrieval models, addressing situations with restricted server access. Dense Retrieval Models involve downloading substantial data volumes in the form of vectorized strips. I conducted extensive literature review and rigorously implemented tools like pyserini and zero-shot retrieval methods. These advancements were integrated into the Lawbot model, significantly boosting retrieval efficiency.Beyond technical skills, this experience cultivated invaluable soft skills such as teamwork, time management, and creative thinking. Adapting to a new continent enriched my growth both professionally and personally.

New York, United States

In a collaboration with Stony Brook University, NY, I contributed to a Natural Language Processing (NLP) research project focused on translating human language text into SQL queries for efficient database information retrieval. We conducted an extensive survey of key NLP frameworks, including Sequence-to-SQL, Type SQL, and Triage SQL, aiming to simplify and unify their complex methodologies into a concise overview.Beyond analysis, we engaged in hands-on work by debugging source code, resolving compatibility issues, and testing these models to ensure their accuracy aligns with documented findings.Our collaborative efforts culminated in the publication of a research paper that not only synthesized the models but also contributed to the broader conversation on NLP-driven database query generation.

Tiruchirappalli, Tamil Nadu, India

I embarked on an intriguing concept that aimed to tackle the significant issue of resource wastage in cloud computing. The project, titled "Cloud Resource Usage Forecasting using Neural Network and Artificial Lizard Search Optimization," was driven by the alarming fact that major cloud providers often allocate surplus resources, resulting in approximately 24% of resource idle time and an annual financial loss of $11 million. To address this challenge, I engaged in a meticulous literature review, identifying the "Artificial Lizard Search Optimization" (ALSO) algorithm as a promising solution. I undertook the task of integrating ALSO as a backpropagation method into a traditional neural network, coding the entire model from scratch in Python to ensure flexibility. A critical aspect of this endeavor involved developing a dynamic lizard allocation approach for optimal efficiency. Subsequently, I evaluated the model's performance through metrics culminating in a research paper.

Tiruchirappalli, Tamil Nadu, India

This project delved into Computer Vision, for the E-commerce sector. The research, titled "DF E-commerce: a deep learning integrated ontology-driven model for enhanced product recommendations," aimed to revolutionize E-commerce recommendation systems.Existing algorithms, often favor either textual or image data. Recognizing the critical role of recommendation systems in E-commerce, I bridged this gap by integrating both modalities. The model combines Machine Learning and Deep Learning techniques to impartially consider user search queries and product images.I implemented stages for intermediate image set generation using a customized web-crawler. The approach leveraged a Dolphin's echolocation metaheuristic algorithm, employing objective functions like SemantoSim, Normalized Google Distance (NGD), and Shannon's entropy to compute semantic similarity.The experiments showcased the superiority of DF E-commerce, achieving 97.3% accuracy with an FDR of 0.04.

Tiruchirappalli, Tamil Nadu, India

During my three-year tenure as Senior Application Developer at Spider, R&D, I led impactful projects. Our team of 20 developers focused on innovation, providing technical expertise to the university, and collaborating with Robotics & Electronics domains to tackle challenges posed by external organizations. One key project involved creating the 2020 Orientation App, vital for freshmen joining during the pandemic. We integrated features such as college maps through Google Maps API and seamless attendance of online orientations via the YouTube API. Using React Native, we ensured cross-platform functionality. The app achieved remarkable success with 1000+ downloads and a 4.9/5 rating.I also played a pivotal role in bringing the Lynx NITT app to life. This app centralized college information, circulars, and club communications, serving as a secure login method for various websites. In my senior year, I mentored juniors, guiding them in development, code refinement, and debugging.

Chennai, Tamil Nadu, India

I interned at Shakti Labs, IIT Madras during the summer of 2021, contributing to the development of the RISC-V system architecture, an open standard instruction set architecture. My focus was on RISCOF, a vital compliance test generator for RISC-V. I worked on RISCV-CTG and RISCV-ISAC, scripting for Bit Manipulation. These engines analyze hardware compatibility, ensuring support for integer, floating-point, and bit operators. Successful tests were crucial, as failure would mean RISC-V couldn't run on that hardware. RISCOF plays a pivotal role in enabling RISC-V's lightweight, dynamic, and versatile operation across devices like tablets, mobiles, laptops, and pagers. My responsibilities included generating bash test cases with RISCV-CTG and executing them using RISCV-ISAC. I also produced coverage reports (certification reports) referencing SPIKE/SAIL as the DUT, expanding my expertise in system architecture and compliance testing.

Kharagpur, West Bengal, India

As part of this internship at Sensor Drops Network, IIT Kharagpur, I contributed to a groundbreaking healthcare project. My role revolved around developing a versatile healthcare application accessible to both doctors and patients. This application was designed for deployment on portable Raspberry Pi devices running Linux.Utilizing Python-Flask within a Docker container, I crafted the application to enable seamless access to patient vitals from anywhere. The application effectively stored time-series data, including ECG, Heart Rate, and SpO2, using InfluxDB and utilized SQLite for biometrics.To ensure data integrity on portable devices, I engineered an automated backup script, triggered periodically or upon network connectivity restoration. Additionally, I played a pivotal role in deploying the application on Raspberry Pi devices, implementing Crontab for efficient scheduling.

Master Of Science - Ms, Artificial Intelligence

Btech., Chemical Engineering

Graduated in First Class with a Distinction and a Minor Degree in Computer Applications.

High School/Secondary Certificate Programs

Achieved 94% in AISSCE examination

Class 10

Achieved 10.0 CGPA without upgradation