Foundation Models Playground

Welcome to the DLAB Foundation Models Playground. This web-based interface allows researchers and developers to interact with state-of-the-art biological science models without writing a single line of code.

Whether you need to predict protein structures, analyze genomic sequences, or calculate molecular properties, the Playground provides a secure, TEE-backed environment to test hypotheses and visualize results instantly.

Accessing the Playground

Navigate to https://playground.delong.lab and log in with your DLAB credentials.

Usage (Upcoming...)

Step 1: Select a Model

Use the Model Hub to choose a model. Models are categorized by domain:

• Protein Structure: (e.g., AlphaFold variants, ESMFold)
• Genomics: (e.g., DNA/RNA language models)
• Molecular Properties: (Toxicity prediction, solubility)
• DeLong Proprietary: Exclusive internal models for specialized tasks.

Tip: Use the AI Search bar at the top to find models using natural language (e.g., "Find models for antibody binding affinity").

Step 2: Input Data

Once a model is selected, the Workspace will adapt to the required input format.

• Text Entry: Paste FASTA sequences or SMILES strings directly.
• File Upload: Upload .pdb, .csv, or .fasta files.
• Load Example: Click "Load Example" to populate the input with valid test data to see how the model works.

Step 3: Configure & Run

1. Adjust Parameters (e.g., recycling iterations for folding models) in the settings panel.
2. Click the Run Inference button.
3. The system will estimate the runtime (e.g., "~45 seconds").

Security Note: Your data is processed inside a Trusted Execution Environment (TEE). No one, including platform administrators, can view your raw input or output data.