90d roadmap · 3 milestones
12-Week Evaluation & AI Reliability
Build an evidence-backed evaluation and observability system for models, tools, and agents. Three four-week milestones move from mechanisms to production trade-offs and a measured synthesis artifact.
Outcome: Build an evidence-backed evaluation and observability system for models, tools, and agents.
How to work this roadmap
Treat each milestone as a claim that must be supported by evidence. Before reading, write what you think the mechanism does and where it will fail. After studying the linked concepts, run the drills without copying an answer, preserve the output, and revise the explanation. Move forward when you can connect the milestone goal to a working implementation, benchmark, architecture decision, or reviewable design artifact.
The 90d horizon is a sequencing aid, not a completion badge. Spend more time where your prediction and the observed behavior disagree. Keep a short decision log containing the mechanism selected, alternatives rejected, expected failure mode, measurement used, and remaining uncertainty. Review that log with FSRS prompts so the roadmap produces durable system judgment rather than a temporary tour of terminology.
At the end, explain Build an evidence-backed evaluation and observability system for models, tools, and agents. from first principles to a reader outside the domain. A strong explanation should survive follow-up questions about correctness, cost, latency, resource use, security, recovery, and operational visibility. If it cannot, return to the milestone that contains the missing mechanism and build a smaller falsifiable example.
Milestones
Milestone 1
Weeks 1-4 — Foundations and mechanisms
Build the domain vocabulary and explain the core mechanisms from first principles.
Concepts
- Structured Outputs
- ML Math Foundations
- Gradient Descent
- Backpropagation
- Softmax & Cross-Entropy
- AdamW Optimizer
- Tokenization (LLM)
- Language Modeling
- Sampling & Decoding
- Embeddings (Transformer)
- Self-Attention
- Multi-Head Attention
- Transformer Block
- Training & Debugging
- LoRA & PEFT
- Probability Fundamentals
- Descriptive Statistics
- LLM Evals
- Model Evaluation
- Coding-agent Benchmarks
- Tool-use Evaluations
Milestone 2
Weeks 5-8 — Production systems and trade-offs
Design the production path, including resource, scale, safety, and operability trade-offs.
Concepts
Milestone 3
Weeks 9-12 — Reliability, verification, and synthesis
Test failure modes, measure outcomes, and ship the synthesis artifact.
Concepts
Build evidence
- Synthesize: Evaluation & AI Reliability — Build an evidence-backed evaluation and observability system for models, tools, and agents. Produce one working system, benchmark, or evidence-backed design that integrates the path.