Unsiloed AI (YC F25) is the unstructured data interface for LLMs and AI agents. We build vision-first, layout-aware systems that use computer vision, OCR, and multimodal models to turn complex documents into deterministic, machine-readable representations.

Role:

We are looking for a passionate and driven Machine Learning Engineer Intern to join our team. This is a research-heavy role, ideal for candidates who thrive on reading and implementing advanced research papers, devising innovative techniques, and working on cutting-edge AI challenges. You will play a pivotal role in building, training, fine-tuning, and deploying multi-modal models tailored for parsing complex unstructured data.

Key Responsibilities

• Conduct in-depth research by reading academic papers and identifying state-of-the-art methods relevant to our domain.

• Build, train, and fine-tune multi-modal models to accurately extract and process information from unstructured data.

• Curate and manage custom datasets, ensuring they are tailored for training and benchmarking models.

• Experiment with and improve state-of-the-art machine learning and deep learning models to achieve higher accuracy and efficiency.

• Benchmark model performance against datasets, focusing on accuracy, efficiency, and robustness.

Requirements

• Currently pursuing or recently completed a degree in Computer Science, Data Science, or a related field with a focus on AI/ML.

• Demonstrated experience in academic research, projects, or publications related to AI/ML, particularly in the areas of multi-modal models and transformer-based architectures.

• Strong programming skills in Python and familiarity with ML frameworks such as TensorFlow or PyTorch.

• Experience with deploying machine learning models, including containerization using Docker and other MLOps tools (e.g., MLFlow, Kubernetes, or similar).

• Knowledge of end-to-end deployment pipelines.

Compensation: 1.5 lakh INR per month for remote role / 6k usd per month for in-office role in SF office.
Location: Remote or in-office

Duration: 2 months initially, extended or converted to full-time role based on performance