| Project Name | Description | Technologies |
|---|---|---|
| Guide on Everything AI | Maintaining a comprehensive guide on everything I've learned regarding artificial intelligence. | Transformers, PEFT, TTS, Latent Diffusion, Fine-tuning techniques |
| Operation Athena | Curating a large database of reasoning tasks that updates in real time and has self moderation capabilities | Next.js, Node.js, MongoDB |
| The Caduceus Project (Code) (Dataset) |
Utilizing GPT-4o to convert highly technical medical and scientific protocls into markdown files. | Python, OpenAI API |
| Pneuma (wip) | Training an LLM on realistic human interactions, conversations, and experiences. | Python, Fine-Tuning, Direct-Preference-Optimization |
| Apocrypha and Sandevistan (wip) | Datasets representing experiences, imaginative scenarios, and other such things we normally don't train LLMs on. | Python, Together.ai API, Prompt-Engineering, Mirroring of Neural Patterns. |
| Study Guide for LLMs (wip) | A dataset to help LLMs study for the ARC-C and MMLU benchmarks by utilizing the Nemotron model. | Python, Nvidia API, Prompt-Engineering |
| System Prompt Generator (unreleased) | A python script that synthetically generates system prompts for ShareGPT datasets | Python, Together.ai API, Prompt-Engineering |
| Interactive Experience Generator (unreleased) | A data pipeline that generates interactions between a human and an AI in ShareGPT format. | Python, Together.ai API, Prompt-Engineering |
| Nemotron 340B Data Generation Pipeline (unreleased) | A data pipeline to generate multiturn data with Nvidia's new 340B Nemotron model, in the ShareGPT format. | Python, Nvidia API, Prompt-Engineering |
| Multiversal Data (wip) | A potential solution to help Large Language Models predict future events. | Python, Together.ai API, Prompt-Engineering |
I'm currently exploring the following technologies:
- Machine Learning
- Cloud Computing
- Data Science
- Deep Learning
Feel free to reach out to me for collaboration, discussions, or just to say hi!
