Darshit is a Machine Learning Engineer with 6+ years of experience specializing in AI technologies and production deployments. He has a strong background in developing and deploying machine learning models, particularly in computer vision and data mining.
Developed and deployed various machine learning models for fraud detection and sound classification.
Automated CI/CD workflows using Python, Jenkins, and GitHub Actions.
Managed production-grade Kubernetes clusters for online casino infrastructure.
Improved ASL sign translation accuracy by integrating text-to-speech capabilities.
Conducted comprehensive data analysis for credit risk assessment.
Successfully developed a classification model to identify 'bad' customers based on credit risk data.
Revamped fraud detection systems with new anomaly detection approaches based on historical gameplay logs.
Successfully developed a machine learning model for accurate classification of violent/non-violent sounds.
Reinforcement Learning for Atari (Space Invaders) — Deep Q Learning agent training with replay memory experimentation.
Key outcomes:
Successfully implemented a Deep Q Learning algorithm to train an Atari game agent.
Improved training results through experimentation with replay memory.
Overview: This project focused on classifying customers as 'bad' or 'good' based on credit risk data. Responsibilities: Conducted comprehensive analysis on customer credit risk data to develop a classification model. Utilized data mining techniques for data collection and pre-processing. Developed and analyzed models using scikit-learn, Matplotlib, and NumPy.
Key outcomes:
Successfully developed a classification model to identify 'bad' customers based on credit risk data.
Overview: This project developed an ASL-to-speech conversion model to benefit the deaf community and caregivers. Responsibilities: Developed and trained the ASL-to-speech conversion model using a custom MediaPipe-generated dataset. Implemented an open-source API for remote and local model access.
Key outcomes:
Successfully developed and trained an ASL-to-speech conversion model.
Overview: Developed a machine learning model to classify sounds as violent or non-violent, crucial for emergency situations. Responsibilities: Developed a machine learning model for accurate sound classification.
Key outcomes:
Successfully developed a machine learning model for accurate classification of violent/non-violent sounds.
Overview: Revamped a fraud detection system by developing new anomaly detection approaches. Responsibilities: Developed new anomaly detection approaches for player gameplay.
Key outcomes:
Successfully revamped an existing fraud detection system.
Darshit
LLM