Jerlin Babu

AI Engineer

• Job Details:Implemented various LLMs (GPT-4, Llama-3, Phi-3.5, Mistral) to automate text generation, correction, and classification, leading to a 40% reduction in manual documentation effort. Extracted and processed large-scale datasets from Hugging Face and Ollama to support model training, increasing model evaluation efficiency by 35%. Integrated multimodal LLMs like Llava, Qwen-VL, and Florence to enable joint text-image-video understanding, improving analytics accuracy by 25%. Developed and deployed classification, regression, clustering, and recommendation systems via FastAPI and MySQL, reducing deployment time by 30%. Utilized LLMs and SLMs for text retrieval and summarization, enhancing content relevance in retrieval systems by 45%. Designed and executed real-time CCTV analytics using YOLOv8 and DeepSORT for object detection and tracking, boosting surveillance accuracy by 50%. Applied advanced deep learning methods for segmentation and feature extraction, resulting in a 20% gain in model precision. Performed data preprocessing and augmentation using Albumentations, improving model robustness by 18%. Enhanced document processing by fine-tuning LLMs and leveraging OCR tools (EasyOCR, PaddleOCR), increasing data extraction accuracy by 30%.

Data Science Intern

• Job Details:Utilized Python libraries for data preprocessing and model deployment, reducing data preparation time by 25%. Managed large-scale data processing using Hadoop & Spark, optimizing data storage and retrieval by 40%. Developed interactive dashboards in Tableau and Power BI, enabling stakeholders to make informed decisions with real-time insights.