NLP Pipeline for Video Generation

• Developed NLP pipeline for generating videos from URLs.
• Implemented a conversation quality metric using Python and FastAPI.
• Deployed solutions via Docker on private Amazon servers.
• Developed an internal user interface dashboard in JavaScript.
• Fine-tuned GPT-neo for conditional text generation.

• Engineered a SOTA multi-document summarization model for news aggregation, achieving an 87% human acceptance rate with less than 1% critical errors.
• Designed automated news timelines for evolving events, resulting in a 2.8% increase in content depth and a 1.4% boost in user time spent.
• Optimized ranking algorithms with a CTR prediction model, increasing daily active users by 4.3% and user engagement by 2.1%.
• Created a novel similarity scoring formula, improving F1-score from 0.91 to 0.95 for news clustering.

• Developed ML Miners on the Bittensor network for Tao and dTao cryptocurrency.
• Managed a team of ML developers on Subnet 44 “Scorevision”.
• Achieved first place consistently by training and optimizing computer vision models.
• Implemented real-time tracking of football players and pitch key points from video footage.
• Trained Yolo models for miners and Clip for validators.

I used vision transformers, data augmention,face detector and deployed the model on AWS SagaMaker

In the corporate sector, text data flows swiftly, emanating from diverse sources in various formats, destined for multiple endpoints and actions. This continuous stream of data necessitates management through structuring, making it easy to interpret and analyse, and ultimately directing it to the appropriate user or applying the insights obtained. I utilised several large language models (LLMs) streamline and enhance this process:

• A question-answering LLM, enabling users to inquire about the content of each document within the flow.
• A summarization LLM, offering a concise overview of each document.
• An NLP model for sentiment analysis, determining the sentiment of each document.
• An NLP model that calculates an absolute similarity score among all documents.
• Clustering, which groups similar documents together and simultaneously identifies outlier documents.