3. Data, Science & AI

Expertly deploys machine learning models via FastAPI, handling model serving and inference endpoints for production deployment.

Generates advertising images with text using Google's Gemini AI model for creative content creation.

Validates data changes for integrity risks including raw edits, column mismatches, and output discipline to prevent system failures.

Analyzes images for content recognition and visual question answering using Zhipu AI's GLM-4V visual language model.

Analyzes chess positions using Stockfish engine, providing best move recommendations and position evaluation via FEN input.

Summarizes K18 QC artifacts by aggregating data and appending manifest rows upon format confirmation.

Executes Quantum ESPRESSO for first-principles DFT, SCF, structural relaxation, band structure, and phonon calculations.

Queries Materials Project, NIST, and PubChem databases for crystal structures, material properties, and thermodynamic data.

Guides review of database changes, migrations, and schema modifications to ensure data integrity and consistency.

Automates AI content generation for affiliate and copywriting platforms, managing prompts, batch processing, and validation.

Expert analysis of warehouse processes using DaRa dataset with REFA methodology, scenario recognition, and validation logic for fact-based insights without hallucinations.

Automates acquisition of computational science resources including force field parameters, pseudopotentials, and crystal structures for research purposes.

Provides nickname-to-formal-name mapping, user validation, and metadata retrieval for consistent user information across report generation skills.

Analyzes simulation data from LAMMPS/QE, computes properties like RDF, MSD, and diffusion coefficients, and generates visualizations.

Retrieves and exports data from Metabase Dev environment, detecting full scan queries and proposing optimization suggestions.

Enables integration with Replicate API to deploy and generate images using AI models like Flux, SDXL, and custom LoRA models.

Runs LAMMPS molecular dynamics simulations for scientific research, including energy minimization, equilibration, and property calculations like diffusion and RDF.

AI-powered tool for merging 2-8 images with Gemini 2.5 Flash guidance via OpenRouter, enabling creative composition.

Analyzes health data to compute health scores and detect abnormal indicators using data science techniques.

Builds conversational AI agents for Next.js frontends using Vercel AI SDK and OpenRouter, supporting streaming UI, tool calling, and multi-provider integration.

Analyzes text for semantic density and compressibility by decomposing into five strata to measure originality and assess conciseness.

Evaluates and tunes the accuracy of the RQ-SCAN model, generating detailed performance reports.

Performs statistical analysis and generates insights from datasets including CSV, Excel, and JSON files for data-driven decision making.

Enables dynamic access to zai-vision MCP server for AI vision tools using stdio transport.