3. Data, Science & AI

Creates custom, interactive data visualizations using D3.js, enabling low-level control for complex charts, maps, and dynamic data binding.

Reviews scientific documents for logical clarity, argument soundness, hypothesis-data alignment, and claims vs evidence to ensure scientific rigor.

Designs neural network architectures respecting user-specified symmetry groups, providing implementation guidance for equivariant models.

Selects optimal data visualizations and generates narrated reports with insights and actionable recommendations for business analytics and data-driven decision making.

Verifies equivariant model correctness by providing symmetry tests and debugging guidance for AI model implementations.

Designs database schemas, models domain entities and relationships, and builds knowledge graphs for structured data representation and system integration.

Enables secure speech-to-text implementation using Faster Whisper, optimizing audio processing and ensuring privacy for voice assistant applications.

Expertly constructs secure prompts, prevents injections, orchestrates multi-step LLM tasks, and validates outputs for the JARVIS AI assistant.

Provides structured scaffolding for data analysis workflows, including TDD, EDA, statistical design, and model validation.

Assists ML engineers in discovering data symmetries and invariances through domain analysis and transformation testing to guide model design.

Establishes statistical baselines using reference classes to anchor forecasts in historical data and avoid cognitive biases.

Statistically optimizes multi-factor systems by identifying key variables, interactions, and response surfaces to maximize performance within constrained experimental budgets.

Maps identified symmetries to mathematical groups (e.g., cyclic, dihedral, Lie) using group theory foundations for architecture design.

Optimizes large language models for deployment via quantization, reducing memory footprint while maintaining performance in resource-constrained environments.

Applies Bayesian inference to update beliefs with evidence, calibrate confidence, and assess risks in uncertain data for improved forecasting and decision-making.

Expert implementation of text-to-speech with Kokoro TTS, covering voice synthesis, audio generation, performance optimization, and secure audio handling for JARVIS.

Expert in graph database schema design, query optimization, and relationship modeling for applications, with specialization in SurrealDB.

Expert guidance on designing high-performance database schemas with normalization, indexing, and full-text search optimization for desktop applications.

Integrates local LLMs via llama.cpp and Ollama with secure model loading, inference optimization, and protection against LLM vulnerabilities.

Enables wake word detection for voice assistants using openWakeWord, featuring audio monitoring, keyword spotting, and privacy-focused always-listening capabilities.

Converts probabilistic forecasts into optimal betting decisions using Kelly Criterion and Brier score optimization for edge calculation and bet sizing.

Evaluates ethical, safety, and fairness impacts of AI/tech decisions, identifies vulnerable populations, and proposes risk mitigation strategies.

Generates images via Google Gemini AI with customizable parameters for creative and technical use cases.

Optimizes and scales cross-model performance for autonomous agent systems, improving efficiency and resource allocation.