Daniel Avila

Build data processing pipelines for machine learning, from preprocessing to batch predictions. Uses Ray Data to handle datasets larger than memory, distribute work across CPUs and GPUs, and feed data into training frameworks.

• Preprocessing datasets larger than memory
• Running batch predictions across multiple GPUs
• Preparing image and tabular data for training

Analyze business data, test hypotheses, and build predictive models with statistical and machine learning methods. Get support across data exploration, experiment design, visualization, result validation, and planning model deployment and monitoring.

• Analyzing customer churn and predicting customers at risk
• Designing an A/B test and evaluating its results
• Forecasting demand for inventory planning

Research markets, competitors, companies, and technical topics using multiple sources in Google NotebookLM. Produce a structured brief with key findings and recommended actions, with optional slides, podcasts, videos, infographics, reports, or mind maps.

• Researching market trends and business opportunities
• Comparing competitors and their positioning
• Researching prospects before a proposal

Get guidance on distributed training with PyTorch FSDP, which splits model parameters across devices. The skill covers implementation, mixed precision, moving data to CPU memory, communication setup, and debugging distributed training code.

• Implement distributed training with FSDP
• Review mixed precision and CPU offloading settings
• Choose a distributed communication backend

Design tests that assess an AI agent's capabilities and reliability. Covers repeated runs to measure variation, checks for required behavior, and adversarial testing to expose failures that ordinary successful examples may miss.

• Measuring variation across repeated runs of the same test
• Checking whether an agent follows required behavior
• Testing responses to unexpected or adversarial inputs

Design AI agents that plan tasks, call tools, and retain useful information. Covers coordination between agents, evaluation, and debugging, with practical patterns for execution limits, error handling, and tracing what an agent does.

• Designing an agent that plans and executes tasks
• Organizing tool definitions and selection
• Deciding what an agent should remember

Manage the limited information an LLM can process at once. Use importance-based summaries, selective trimming, token counting, and content placement to plan what stays in context and address information loss during long conversations.

• Planning how to handle conversations near the token limit
• Summarizing a long dialogue while preserving key information
• Reviewing the placement of important instructions and facts

Automatically retrieves current library and framework documentation through the Context7 API during Claude Code conversations. Use it when asking about tools such as React, Next.js, or Prisma and needing up-to-date reference material.

• Retrieve current React documentation during a conversation
• Consult Next.js documentation in Claude Code
• Fetch Prisma reference material when asking questions

Design memory systems that help AI assistants retain conversation information and retrieve it when relevant. Explore short-term, long-term, and entity memory, along with memory consolidation, prompt integration, storage lifecycle management, and separation between users.

• Designing an assistant that remembers across conversations
• Storing and updating facts about people or other entities
• Adding relevant memories to prompts

Investigate production issues by searching Datadog logs, querying metrics, and following request traces across services. Compare error counts across time periods, inspect surrounding logs, and manage dashboards through a command-line tool.

• Investigating errors in production
• Comparing error counts across time periods
• Following request traces across services

Design collaborative AI teams with CrewAI by defining each agent’s role, goals, tasks, and expected results. Covers passing work between agents, manager-led coordination, memory settings, and execution planning for workflows such as research and writing.

• Build a team that researches and writes articles
• Coordinate specialists through a manager agent
• Define agent roles and expected task outputs

Helps speed up LLM text generation by comparing methods that predict and verify multiple tokens at once. Covers draft model selection, Medusa prediction heads, Lookahead Decoding settings, and deployment examples for inference services.

• Reducing chatbot response latency
• Comparing methods for faster code generation
• Selecting a small draft model for a larger LLM

Build and train Mixture of Experts models, which activate selected specialist networks for each input. Use DeepSpeed or HuggingFace workflows to configure routing, balance workloads, distribute experts across GPUs, and optimize inference.

• Designing MoE models for a limited compute budget
• Implementing Mixtral-style expert routing
• Adjusting expert workloads and training stability

Interact with Langfuse and access its documentation. Use when needing to (1) query or modify Langfuse data programmatically via the CLI — traces, prompts, datasets, scores, sessions, and any other API resource, (2) look up Langfuse documentation, concepts, integration guides, or SDK usage, or (3) understand how any Langfuse feature works. This skill covers CLI-based API access (via npx) and multiple documentation retrieval methods.

The English summary is being prepared. Showing the original description.

Write a pull request description with summary, motivation, and test plan. Use when creating a PR, asked for a PR description, or preparing to submit a pull request.

The English summary is being prepared. Showing the original description.

Write an Architecture Decision Record (ADR) document for a technical decision. Use when documenting a technical choice, architectural decision, or design decision.

The English summary is being prepared. Showing the original description.

Research a technical topic thoroughly by exploring the codebase and summarizing findings. Use when asked to investigate how something works, research an area of the codebase, trace a data flow, or understand a component in depth.

The English summary is being prepared. Showing the original description.

Explains code in plain language for someone unfamiliar with the programming language. Use when asked to explain code, walk through logic, describe what a function does, or when the user says "explain this" or "walk me through this".

The English summary is being prepared. Showing the original description.

Extract text and tables from PDFs, fill PDF form fields programmatically, and merge multiple PDF files into one. Use when working with PDF documents, form filling, document extraction, splitting, or any PDF manipulation task.

The English summary is being prepared. Showing the original description.

Analyze the current uncommitted diff and give a detailed summary with risk assessment. Use when asking what changed, reviewing changes before a commit, or requesting a change summary.

The English summary is being prepared. Showing the original description.

Perform a deep architectural analysis or security review requiring extended reasoning. Use for complex architecture decisions, security audits, or when standard analysis is not finding the root cause.

The English summary is being prepared. Showing the original description.

Deploy the application to production. Only invoke this manually when you are ready to deploy.

The English summary is being prepared. Showing the original description.

Generate a conventional commit message for the current staged git changes. Use when the user wants to commit, asks for a commit message, or asks what to write for a commit.

The English summary is being prepared. Showing the original description.

Lists the conventions for this project and demonstrates the project scope for Claude Code skills. Use when asked about project conventions, code style, or as a demonstration of project-scoped skills.

The English summary is being prepared. Showing the original description.