Saurabh S.

Key Skills

Software

CVAT

Top Subject Matter

Automated Driving Test System (driving license testing)

RAG-based intelligent text generation and question answering

Web crawling for RAG/NLP data collection

Top Data Types

Image

Text

Audio

Top Task Types

Object Detection

Question Answering

Data Collection

Freelancer Overview

Trainee Engineer - IBM Watson Chatbot and CV/annotation work (Project: ADTS driving license automation). Brings 6+ years of professional experience across complex professional workflows, research, and quality-focused execution. Core strengths include CVAT, Internal, and Proprietary Tooling. Education includes Bachelor of Engineering, Government Engineering College, Modasa (2021) and Higher Secondary Certificate, Vivekanand High School, Ahmedabad (2017). AI-training focus includes data types such as Image, Text, and Audio and labeling workflows including Object Detection, Question Answering, and Data Collection.

Labeling Experience

AI/Computer Vision Engineer - Silver Touch Technologies

CVATImageObject Detection

Developed a computer vision automation system for driving license testing using object detection. Trained and fine-tuned YOLO-based and CNN models using custom datasets to improve detection reliability. Integrated caching and in-memory storage to reduce database load and support responsive real-time evaluation. • Built YOLO-based object detection models for driving test automation • Fine-tuned custom CNN models using CVAT and Roboflow annotation workflows • Implemented Redis for faster in-memory handling to improve runtime performance • Delivered an end-to-end pipeline integrating model inference with application logic.

2022 - Present

AI/ML Software Engineer - Silver Touch Technologies

CVATImage

Engineered AI/ML solutions with an emphasis on retrieval-augmented generation and computer vision for production use cases. Built robust pipelines that combine data crawling, embedding-based retrieval, and instruction-tuned language models for context-aware responses. Applied engineering practices to optimize performance, accuracy, and real-time behavior of downstream systems. • Developed an end-to-end RAG pipeline for automated intelligent text generation • Implemented fast web crawling with Crawlee for dynamic content ingestion • Built embedding generation and similarity search for efficient retrieval • Performed prompt engineering and integrated LLaMA/Dolly for accurate response generation.

2022 - Present

AI/ML Software Engineer (Project: BotGenie)

TextQuestion AnsweringData Collection

Developed an end-to-end Retrieval-Augmented Generation (RAG) system for automated intelligent text generation and context-aware responses. Built embedding generation and similarity search components to retrieve relevant context for downstream instruction-tuned language models. Performed prompt engineering and integrated open-source LLMs for accurate answer generation. • Implemented RAG pipeline for automated intelligent text generation.• Used embedding generation and similarity search for context retrieval.• Performed prompt engineering to guide instruction-tuned models.• Integrated open-source LLaMA and Dolly to generate context-aware responses.

2022 - Present

Trainee Software Engineer - Silver Touch Technologies

CVATAudio

Worked as a trainee engineer supporting chatbot development and AI-enabled application features. Built a Watson-based conversational experience using user prompts and website content as knowledge inputs. Implemented speech-to-text and text-to-speech capabilities and automated web crawling and text preparation for question answering. • Developed an IBM Watson chatbot using prompts and website content • Annotated and prepared text datasets and integrated web crawling for QA • Implemented Slack speech-to-text and text-to-speech using Watson APIs • Supported integration and iteration of conversational features based on testing feedback.

2021 - 2022

Trainee Engineer - IBM Watson Chatbot and CV/annotation work (Project: ADTS driving license automation)

CVATImageObject Detection

Worked on a computer vision pipeline for automated driving license tests using YOLO-based object detection models trained on custom datasets. Contributed to dataset labeling workflows by fine-tuning custom CNN models with annotation tooling support from CVAT and Roboflow. Focused on producing accurate detections through model training on annotated imagery.• Built and fine-tuned YOLO object detection for driving test automation.• Used CVAT and Roboflow for image annotation to create training data.• Trained custom CNN architectures on the annotated dataset.• Integrated results into a test automation system for real-time performance.

2021 - 2022

Education

Work History