Suraj R.

Key Skills

Software

Other

Top Subject Matter

Medical computer vision and text processing

Regulatory document text processing and LLM-based contradiction detection

Top Data Types

Image

Text

Medical Dicom

Top Task Types

Text Generation

Segmentation

Classification

Entity (NER) Classification

Text Summarization

Fine-tuning

Freelancer Overview

Graduate Researcher — AI Agents and Foundation Models for Predictive Embryology in IVF Procedures (Weill Cornell Medical. Brings 8+ years of professional experience across complex professional workflows, research, and quality-focused execution. Core strengths include PyTorch, TensorFlow, and Other. Education includes Doctor of Philosophy, Weill Cornell Medical College (2026) and Bachelor of Science, Georgia Institute of Technology (2021). AI-training focus includes data types such as Medical and DICOM and labeling workflows including Evaluation, Rating, and Text Generation.

Labeling Experience

Graduate Researcher — AI Agents and Foundation Models for Predictive Embryology in IVF Procedures (Weill Cornell Medical College)

Developed and trained an embryo foundation model to support real-time embryo monitoring and grading in IVF. The work focused on learning from a large image dataset and using the model for clinically relevant prediction outputs. The system was integrated into clinical workflows to improve clinician efficiency and support continuous monitoring. • Engineered the FEMI masked autoencoder foundation model on 20 million embryo images. • Built conditional diffusion models to generate synthetic embryo images for dataset augmentation. • Produced predictions for embryo quality and ploidy assessment using the foundation model. • Implemented AI agents to grade and monitor embryos 24/7 in the clinic workflow.

2021 - Present

Data Scientist Intern — Streamlining Business Processes using AI and LLMs (Regeneron)

OtherText Generation

Built NLP/LLM-based analytics to process unstructured text for business and regulatory document tasks. The activity emphasized using language models to classify and detect contradictions within regulatory content. This included transforming raw text into structured, actionable outputs using GPT-style tooling. • Reduced unstructured text processing time by 95% using NLP and LLMs for categorization. • Led a contradiction detection project for regulatory documents using a comparison model to improve integrity by 80%. • Implemented analytics and data visualization tools using GPT-4 and pandas AI to create actionable insights. • Operationalized language model outputs to support downstream interpretation of regulatory text.

2024 - 2024

Education

Work History