skills.rest
Tgoldi
TgoldiCOMMUNITY
Explore all skills by this author →
GitHub

senior-ml-engineer

Deploy ML models, build MLOps pipelines, and integrate LLMs with RAG systems.

1|Updated May 10, 2026
Search Tags:#drift detection#mlops#rag#model deployment#llm integration#feature store#model monitoring
One-click install
npx skills add https://github.com/Tgoldi/claude-skills --skill senior-ml-engineer-tgoldi
Or copy as Structured Prompt for Agent▼
Please help me install this Agent Skill.
Skill: senior-ml-engineer
Source: https://github.com/Tgoldi/claude-skills/tree/main/senior-ml-engineer
Command: npx skills add https://github.com/Tgoldi/claude-skills --skill senior-ml-engineer-tgoldi

SYSTEM DOCUMENTATION & REQUIREMENTS

💡 This Skill includes scripts (resource) and references (resource) components.

What problem does it solve? Moving machine learning models from notebooks to production requires deployment infrastructure, monitoring, and LLM integration patterns that most teams build from scratch. This Skill provides production-ready workflows for model serving, MLOps pipelines, RAG systems, and drift monitoring. ## Core Features & Use Cases - Model Deployment: Containerized serving with FastAPI, Triton, or TorchServe, including canary releases and Kubernetes manifests. - MLOps Pipelines: Feature stores with Feast, experiment tracking with MLflow, A/B testing infrastructure, and automated retraining triggers. - LLM & RAG Integration: Provider abstraction layers, retry/fallback logic, cost tracking, vector database selection, chunking strategies, and reranking. - Use Case: You need to deploy a churn prediction model with drift monitoring. Use this Skill to generate the Dockerfile, Kubernetes deployment, and KS-test drift detection, then set alerts when PSI exceeds 0.2. ## Quick Start Ask the AI to help you deploy a trained model to production with monitoring, or to scaffold a RAG pipeline with a vector database and reranking.

Frequently Asked Questions about senior-ml-engineer

High-intent search queries and answers about installing and using this skill.

FAQPage Schema
How do I deploy a machine learning model to production?▼

Export the model to ONNX or TorchScript, package it in a Docker container with dependencies, deploy to staging, then release a canary to 5% of traffic. Promote to full production once p95 latency stays under 100ms and error rate under 0.1%.

Which vector database should I use for a RAG system?▼

Pinecone suits managed production workloads, Qdrant offers the lowest latency for performance-critical cases, and pgvector works well if you already run PostgreSQL. Chroma is appropriate for prototyping before committing to infrastructure.

How do I detect data drift in production ML models?▼

Use the Kolmogorov-Smirnov test to compare reference and current feature distributions, flagging drift when the p-value drops below 0.05. For alerting, treat PSI above 0.1 as a warning and above 0.2 as critical.

How do I handle LLM API rate limits and failures?▼

Wrap calls with exponential backoff retry using tenacity, stopping after three attempts. Implement a provider abstraction layer so you can fall back to a secondary provider when the primary fails repeatedly.

What triggers automated model retraining?▼

Common triggers include scheduled cron jobs, accuracy dropping below a threshold, data drift with PSI above 0.2, or sufficient new data volume for incremental updates. Each trigger runs the training pipeline and validates the new model against the production baseline.

Related Skills

Explore More Software Engineering Skills→
raoxuan98-hashraoxuan98-hash
2

fuxictr-guide

Guide FuxiCTR configuration for CTR model training with run_expid.py.

Community
Intermediate
Proxy2021Proxy2021
5

api_monitor

Aggregate endpoint metrics, latency, and alert status to monitor API health.

Community
Advanced
coreymaypraycoreymaypray
1

model-qa

Audit AI models with evaluation, calibration, and SHAP analysis.

Community
Advanced
davila7davila7
30.5k

ml-engineer

Implements production ML systems with PyTorch, TensorFlow, model serving, and monitoring.

Community
Intermediate