ABHISHEK SINGH

Machine Learning Engineer | AI Developer

Greater Noida, IN.

About

Highly motivated and results-oriented Machine Learning Engineer and AI Developer with a strong foundation in deep learning, computer vision, and full-stack application development. Proven ability to design, implement, and optimize complex ML models, enhance system modularity, and contribute to open-source projects. Eager to leverage expertise in PyTorch, Python, and advanced ML techniques to drive innovation in challenging technical environments.

Work

Lightning AI

|

Open-Source Contributor

→

Summary

Contributed to the Lightning AI open-source project, focusing on improving code efficiency, modularity, and addressing critical bugs to enhance framework stability and performance.

Highlights

Refactored core utility functions to enable conditional NumPy imports, enhancing adaptability and resource efficiency across the framework.

Improved code modularity by localizing imports and minimizing global dependencies, significantly boosting maintainability and reducing complexity.

Resolved a critical bug related to Distributed Data Parallel (DDP) alias usage, ensuring stable and efficient distributed training operations for users.

Defense Research and Development Organisation (DRDO)

|

Machine Learning Intern

→

Summary

Developed and implemented a machine learning solution for vehicle classification, focusing on deep learning techniques and data preprocessing for robust performance.

Highlights

Developed and trained a Convolutional Neural Network (CNN) for military vehicle classification, leveraging transfer learning with the VGG16 architecture to achieve high accuracy.

Preprocessed extensive datasets of military vehicles and utilized VGG16 for efficient feature extraction, integrating custom dense layers to optimize classification performance.

Gained hands-on experience with Python, PyTorch, Keras, VGG16, ImageDataGenerator, and Matplotlib in a real-world defense research environment.

Education

KCC Institute of Technology and Management

→
B.Tech Computer Science and Engineering Grade:

Skills

Programming Languages

Python.

Machine Learning Frameworks

PyTorch, Keras.

Computer Vision Libraries

OpenCV, scikit-image.

Data Manipulation & Visualization

NumPy, Matplotlib, Seaborn, ImageDataGenerator.

Deep Learning Architectures

VGG16, U-Net, Generative Adversarial Networks (GAN), Convolutional Neural Networks (CNN).

Machine Learning Concepts

Machine Learning, Deep Learning, Computer Vision, Transfer Learning, Semantic Segmentation, Image Enhancement, Large Language Models (LLM), AI Agents.

Evaluation Metrics

IoU (Intersection over Union), Dice Coefficient, Pixel Accuracy, PSNR (Peak Signal-to-Noise Ratio), SSIM (Structural Similarity Index Measure), Confusion Matrix, ROC Curve, Precision-Recall Curve.

Web Technologies

React, FastAPI, Full-Stack Development, Web Development.

Specialized Techniques

Adversarial Loss, L1 Loss, Perceptual Loss, Medical Imaging, Distributed Data Parallel (DDP).

Projects

WorldForge AI: An AI Agent for Fictional Genre Writers

→

Summary

Led the engineering of a full-stack AI application leveraging large language models for dynamic fantasy world generation and immersive narrative exploration.

GAN-Based Endoscopic Image Enhancement ML Model

→

Summary

Engineered a PyTorch-based Generative Adversarial Network (GAN) to enhance low-quality endoscopy video frames, improving diagnostic clarity.

Pixel level Face Segmentation ML Model

→

Summary

Developed a pixel-level face segmentation model using U-Net for semantic segmentation, focusing on optimizing performance and evaluation metrics.