Parikshit Kumar | ML Engineer

About Me

I enjoy building things that work—especially when they involve data, code, and a bit of creativity. My background is rooted in machine learning, generative AI, and computer vision, with hands-on experience in both research and engineering. I value clarity, thoughtful design, and practical solutions. Whether collaborating on a new idea or refining a complex system, I bring focus, curiosity, and a steady approach to my work.

Skills

Technical Skills

Deep Learning

NLP

LLMs

Computer Vision

Generative Modeling

Diffusion Models

Machine Learning

Reinforcement Learning

Data Analysis

Data Visualization

Statistics

Graph Neural Networks

Platforms & Tools

Python

C++

Keras

TensorFlow

PyTorch

PyTorch Lightning

FastAI

OpenCV

Numpy

Matplotlib

Pandas

Hugging Face

SQL

Flask

BeautifulSoup

XGBoost

Tableau

MLOps

GCP

AWS

Kubernetes

Docker

Git CLI

Jupyter Notebooks

Anaconda

Linux Terminal

ROS

CUDA

MPS (Apple)

Apache Spark

Tensorboard

Firebase

MS Azure

Projects & Papers

Generative Networks and VAEs to Assist Artists - SRM IST, Paper and Project
Apr 2021 – Jun 2021

Designed Variational Autoencoders (VAEs) for sketch refinement, improving workflow efficiency by 30% and enhancing creative control.
Achieved 90.6% accuracy with the CUHK Face Sketch Database, presenting findings at ICCET 2023 in India.
Implemented advanced techniques to retain finer details in generated images, significantly improving visual fidelity.
Built a user-friendly interface for artists to seamlessly integrate the model into their creative workflows.
Validated real-world usability with artists, receiving positive feedback on practicality and output quality.

Python Keras TensorFlow OpenCV VAE CUHK Face Sketch DB Generative Models ICCET 2023 Paper

Energy Price Forecasting Using LSTMs and ARIMA - SATU, Paper and Project
Apr 2021 – Apr 2021

Developed LSTM and ARIMA models for energy price forecasting, improving prediction accuracy by 18% for Australian energy datasets.
Conducted NLP-based trend analysis to identify hidden patterns, providing actionable insights for stakeholders.
Improved model training speed through data preprocessing techniques, enabling real-time forecasting capabilities.
Built visual dashboards using Tableau, simplifying interpretation for non-technical decision-makers.
Presented findings at the AIJR Conference 2022, highlighting the balance between model performance and scalability.

Python LSTM ARIMA NLP Tableau Pandas Data Analysis AIJR Conference 2022 Paper

Personal Fine-Tuned LLM Assistant - UC Riverside, 2024
2025

Created a personal LLM application fine-tuned on a Final Fantasy character dataset for daily assistance.
Allows users to select their favorite character as their AI helper for a personalized experience.
Utilizes the totl llama framework for efficient model serving and interaction.
Supports context-aware responses and task automation tailored to user preferences.
Designed for privacy, running locally without cloud dependencies.

Python LLM totl llama Final Fantasy Personal Assistant NLP UC Riverside 2024 2025

FundPilot: AI-Powered Web Form Automation
2025

Automates web form filling using AI models and Selenium-based web automation tools.
Generates context-aware answers for form fields using advanced language models.
Designed for offline operation to maximize privacy and efficiency.
Integrates Python 3.x, Chrome, ChromeDriver, and Ollama with LLaMA 2 7B for robust automation.
Ensures compatibility and easy setup across different environments and browser versions.

Python Selenium Ollama LLaMA 2 7B ChromeDriver AI Automation 2025

Anime Aesthetic Transformation with Pix2Pix GAN
Apr 2021 – Jun 2021

Utilized Pix2Pix GAN for transforming traditional anime images into enhanced, high-quality artworks.
Trained the model on a diverse dataset of anime images, achieving superior color correction and detail preservation.
Implemented an iterative feedback mechanism to refine output aesthetics based on user suggestions.
Presented the project at a local tech fair, receiving commendations for blending technology and creativity.
Optimized data preprocessing pipelines, reducing GPU training time by 20% without compromising output quality.

Python Pix2Pix GAN Deep Learning OpenCV Data Preprocessing 2021

Starterdex+ Pokémon Identifier
2021

Built an app to identify all 151 Gen 1 Pokémon and 15 mega evolutions using DenseNet201 and FastAI over PyTorch.
Provided detailed information and stats for each identified Pokémon.
Utilized advanced image classification for high accuracy and reliability.
Designed a modern, user-friendly interface for quick Pokémon lookup.
Optimized for mobile and desktop use, ensuring accessibility for all users.

Python DenseNet201 FastAI PyTorch Pokémon Image Classification 2021

Wasserstein GAN Art Generator
2021

Developed a web application based on Wasserstein GAN with Gradient Penalty (WGAN-GP) to generate realistic paintings.
Trained the model on a dataset of 8,000 Albrecht Dürer paintings for high-fidelity artistic output.
Implemented a super-resolution mode to enhance generated images for detailed viewing.
Designed an intuitive user interface for generating and downloading artworks.
Optimized model performance for real-time generation and minimal latency in the web app.

Python WGAN-GP TensorFlow Super-Resolution Web App Art Generation 2021

Hand Gesture Recognition via Webcam
2020

Developed a real-time hand gesture recognition system using a standard computer webcam.
Implemented convolutional neural networks (CNNs) for accurate gesture classification.
Enabled sign language recognition and automation tasks based on detected gestures.
Optimized the model for low-latency inference and robust performance in varied lighting.
Designed a user-friendly interface for live gesture feedback and system control.

Python CNN OpenCV Webcam Sign Language Automation 2020

Optimized Firearm Detection
Dec 2020 – Feb 2021

Developed a firearm detection application using OpenCV and YOLOv3, achieving high precision and recall metrics.
Trained on a custom dataset of images and videos, optimizing the model for real-time security applications.
Integrated the model with CCTV feeds for real-time detection, enabling proactive security monitoring.
Optimized inference time through quantization and efficient model pruning, achieving a 25% reduction in latency.
Tested the model under varying lighting conditions and environmental noise to ensure robust performance.

Python YOLOv3 OpenCV Deep Learning Computer Vision Quantization Model Pruning 2020

Experience

Junior Specialist - University of California, Riverside

Feb 2025 – Dec 2025

Engineered deep learning models for mosquito orientation classification, achieving 96% accuracy across four orientations.
Utilized advanced image processing techniques—including thresholding, edge detection, and morphological transformations—for robust biological feature extraction.
Integrated model deployment pipelines with interdisciplinary teams, optimizing end-to-end system performance and scalability.
Enhanced scalability of feature extraction pipelines, ensuring adaptability for future biological studies.
Documented research findings and contributed to publications, sharing insights with the scientific community.

Student Assistant, Computer Vision Research - University of California, Riverside

Jul 2024 – Dec 2024

Conducted sensitivity analysis with increased sample sizes, reducing variance by 20% and improving detection accuracy by 30%.
Collaborated with entomologists to optimize algorithms tailored to specific biological requirements.
Analyzed mosquito structures in detail using advanced image processing for biological feature extraction.
Developed preprocessing pipelines to streamline data preparation for deep learning models.
Provided technical support and training to research team members on computer vision tools and workflows.

Research Assistant - University of Guelph, Ontario

Jan 2023 – Jun 2023

Utilized image segmentation with LLM-based text analysis to assess soil properties, improving accuracy by 18%.
Designed an automated application, reducing data processing time by 40% and enhancing indicator detection accuracy by 25%.
Collaborated with soil scientists to refine model parameters for precise results.
Integrated feedback from cross-disciplinary teams to ensure model usability for practical applications.
Co-authored a paper on data-driven soil characterization, contributing to a 15% performance boost.

Computer Vision Intern - UT Dallas & IIT Bombay, Remote

Mar 2022 – Jul 2022

Created deep learning algorithms for autonomous vehicles, boosting object detection accuracy by 15% using Faster R-CNN and YOLO.
Enhanced segmentation accuracy by 20% with 3D simulations, improving off-road vehicle performance.
Developed a preprocessing pipeline, reducing training time by 25% for large datasets.
Collaborated with a multidisciplinary team to refine detection algorithms for varied terrain types.
Integrated machine learning models into real-time applications, optimizing system responsiveness.

Machine Learning Intern - Indian Institute of Technology, Bombay

Jan 2021 – Sept 2022

Built knowledge curation engines for large datasets by applying ML/NLP techniques, optimizing pipelines to improve search relevancy and information retrieval.
Analyzed unstructured data to develop classification models, leveraging feature engineering and hyperparameter tuning to enhance model accuracy.
Designed clustering algorithms for dataset categorization, improving downstream task efficiency and reducing data redundancy.
Automated anomaly detection in large-scale datasets, achieving a 20% reduction in manual analysis overhead and streamlining data verification processes.
Optimized indexing strategies in data pipelines, reducing query response times by 35% and improving system performance for real-time queries.