Himanshu Nimonkar

Master of Science in Computer Science

University of California, Davis

GPA: 4.00/4.00

Bachelor of Technology in Information Technology (with Honors in Artificial Intelligence)

K. J. Somaiya College of Engineering, Mumbai

GPA: 3.81/4.00

$1,500 Scholarships

Top 5% of the Class

Graduate Research Assistant (Software Engineering Team)
Dr. Dongyu Liu’s Visualization and Intelligence Augmentation (VIA) Lab, UC Davis

• Researched Context-Aware Prediction methodologies on time-series data and built a comprehensive pipeline software that integrates multiple worker model architectures (monolithic and mixture configurations) across datasets of various domains. Developed a unified data processing pipeline with direct CSV-to-tensor conversion, lazy loading capabilities for GPU memory optimization, and eliminated intermediate file dependencies. Established performance benchmarks and implemented Optuna-based hyperparameter optimization for model tuning and validation.

• Ideated Sintel/Orion's anomaly detection framework's refactor with a modular design and optimized data transformations, reducing computational overhead by 9%.

Graduate Research Assistant (Machine Learning Team)
Department of Wildlife, Fish and Conservation Biology, UC Davis & School of Renewable Natural Resources (College of Agriculture), Louisiana State University

• Built CNN-based retraining pipeline for anthropogenic disturbance in bioacoustics research, processing 720+ hours of ARU audio data across 30 wetland sites, parsing RavenPro spectrograms and aligning .flac files via timestamp parsing algorithms, thereby improving hunter-produced gunshot detection by 8%. Set up a reproducible Conda environment with GUI deployment.

Software Engineering Intern
DeepCytes Cyber Labs (UK)

• Engineered an Android security testing app with Kotlin to find permission vulnerabilities in system functions (location, camera, mic, SMS, media), reducing security incidents by 42% in 5 enterprises, while leveraging Firebase for secure data management.

• Built real-time CVE/CWE data tracker with auto email alerts via user watchlists for network monitoring/encryption. Added host file tweaks boosting malware defense by 29%. Optimized deployments using GitHub Actions, AWS EC2, Nginx and Cloudflare.

• Built a CLI-based red-teaming automation toolkit to stress test apps and model threats, which detected 37 unidentified attack vectors with Burp Suite, Reconnaissance, Maltego, Shodan, BeEF, Metasploit, Nessus, John the Ripper, and Nmap to remediate vulnerabilities. Implemented CI/CD on AWS EC2 using Ansible playbooks and Jenkins, reducing deployment time by 8%.

• Architected DarkwebGPT, a threat intelligence tool using the DarkBERT API and Agile methods for dark web analysis.

• Built a LangChain-based legal assistant with Hugging Face transformers for IPC queries, increasing engagement by 25% while implementing QA models and LLMs with a Gradio interface. Set up a Docker pipeline on AWS EC2 to deploy MySQL databases for scaled data operations. Developed RESTful CRUD API based dashboards using postman, lifting user retention by 41%.

• Built test automation suite (UI, unit, integration) for DC website with Jest and Selenium, cutting bug detection time by 35%.

• Built an OSINT framework for cyber fraud analysis, improving case response time by 15% across the UK, EU, and APAC regions.

• Optimized deepfake detection via a hybrid ResNeXt CNN-LSTM architecture, improving accuracy by 9% for media forensics.

Machine Learning Research Intern
Dr. Irfan Siddavatam and Dr. Ashwini Dalvi Lab, K. J. Somaiya College of Engineering

• Published ”A Transfer Learning Approach for Classification of Knee Osteoarthritis” (ICEEICT 2023, IEEE Xplore), achieving 82.5% accuracy and 97.62% precision with a transfer-learned DenseNet-201 CNN model trained on 5,748 X-ray images for early OA detection with AI. Awarded BEST PAPER in technical session CS 03. DOI: 10.1109/ICEEICT56924.2023.10157147

• Published ”User Reception is Everything: Using a Neural Network to Predict iOS App Ratings” (ICDSA 2023, Springer LNNS), implementing ELECTRA transformer architecture on 97 multi-genre apps’ reviews, achieving 73% accuracy and 0.705 F1 score, with seaborn, sklearn, PyTorch, and NLTK for multi-class rating prediction. DOI: 10.1007/978-981-99-7820-5 34.

• Built air canvas framework for gesture-based character recognition, gaining 88.79% accuracy with TensorFlow and Mediapipe.

Information Interfaces Research Intern
Dr. Irfan Siddavatam Lab, Somaiya School of Design

• Published four IEEE and Springer papers on user-centered design for educational assessment, mental health support, pet welfare tracking, and university opportunity management systems. Research focused on identifying user needs through literature reviews and surveys, developing tailored interfaces, and validating designs through user testing.

Software Development Intern
The Marine Robotics Team, KJSCE

• Contributed to the development of YOLO v4-tiny and v5-based image processing models for detection tasks (gates, flares, G-man, Bootlegger, and underwater trash), and developed ROS Melodic Gazebo simulations for our AUV, Proton.

Web Development Intern
Sparkling Wings NGO

• Engineered WordPress website with domain configuration while integrating Google Maps API and Stripe API.

Software Testing Intern
Reha App

• Detected bugs in the software, including error handling and UI bugs.

Joint Sports Secretary

Managed sports & cultural festivals (20k+ attendees) at KJSCE.

Treasurer, Emfinity Math Club

Secured sponsorships, managed finances, and organized campus math events.

Vice Captain, The Marine Robotics Team

Led a multifaceted 22-member team in developing autonomous underwater vehicles. Also handled the budget, logistics, and permissions for the team.

Sparkling Wings NGO

Taught English & Math weekly to underprivileged children.

KJSCE Parvaah

Organized yoga day and blood donation drives. Worked with BMC (Mumbai Government) to clean Mithi River and Juhu Beach

Agasthya International Foundation

Mentored 6 school students of Raje Sambhaji School (Khar) to build an NFC-based women's safety hardware box as part of the Design Thinking Challenge organized by JP Morgan Chase & Co.