A daily roadmap for becoming dangerous at AI engineering.

Start with the mechanics behind modern language models from an engineering lens: tokens, model families, sampling, streaming, structured outputs, tool use, multimodal inputs, embeddings, and reasoning models.

Learn prompt design as software design: message roles, few-shot examples, persona boundaries, templating, output parsers, and prompt injection defense.

Context engineering is the discipline of managing system prompts, conversation history, memory, tool results, retrieval, and compression under real context-window limits.

Retrieval augmented generation connects models to private knowledge. Learn chunking, vector databases, hybrid search, reranking, metadata filters, GraphRAG, document parsing, and multimodal retrieval.

Move beyond single prompts into systems that plan, act, observe, delegate, retry, and preserve state across long-running tasks.

Tool use turns a model from a text generator into an operator. Learn schema design, parallel tool calls, MCP, sandboxed execution, API wrappers, and tool result formatting.

Inference engineering is where model quality meets infrastructure reality: hosted APIs, self-hosting, vLLM, TGI, SGLang, Ollama, llama.cpp, quantization, caches, batching, and cost-performance trade-offs.

Production AI needs traces, token accounting, model and prompt versions, A/B tests, replay tools, drift detection, and feedback loops.

Evaluation engineering turns subjective AI behavior into testable product quality using golden datasets, LLM judges, RAG metrics, agent trajectory analysis, synthetic data, online evals, and red teaming.

AI product margins depend on routing, caching, batching, distillation, compression, token budgets, and model gateways.

AI security is not a checkbox. Learn prompt injection, jailbreak resistance, tool-call exfiltration, moderation, guardrail frameworks, PII redaction, abuse prevention, schema validation, and audit logging.

Multimodal systems combine text with vision, audio, documents, generated images, video, and realtime experiences. Learn the capabilities and engineering constraints before building shiny chaos.

AI app architecture connects model calls to product systems: streaming, background jobs, resumable agent runs, tenant-level key management, fallback chains, feature flags, and APIs for cancel/retry/partial results.