# Certifications — an honest, tiered guide

> **Read this first.** For an **AI/LLM engineering** role, a **shipped RAG/agent project on GitHub** (exactly what this plan builds) beats any certificate. The research is blunt: the strongest AI candidates tend to hold the *fewest* certs, and a stack of "prompt engineering" badges can read as a **negative** signal. So treat certs as a **complement** — an ATS/recruiter-screen passer and a tiebreaker — not the main event. **Do the project; add one or two certs around it.**
>
> All links verified live (2026).

## Tier 1 — Worth‑it recognized exams (pick **one**, near the end)

Proctored industry exams recruiters actually recognize. Do these *after* you've built the product, as a screen‑passing tiebreaker.

| Cert | Cost | Fit | Notes |
|---|---|---|---|
| **[Databricks Certified GenAI Engineer Associate](https://www.databricks.com/learn/certification/genai-engineer-associate)** | $200 | ⭐ **high** | **The best topical match** — RAG, chunking, embeddings, vector search, agents, evals. If you do one, do this. (Some Databricks‑stack specifics.) |
| **[AWS ML Engineer – Associate (MLA‑C01)](https://aws.amazon.com/certification/certified-machine-learning-engineer-associate/)** | $150 | high (AWS) | Strong ATS keywords ("AWS"+"ML"); fintech skews AWS. More MLOps than LLM. |
| **[Azure AI Engineer Associate (AI‑102)](https://learn.microsoft.com/en-us/credentials/certifications/azure-ai-engineer/)** | $165 | high (Azure) | Refreshed with GenAI/agents/RAG on Azure. Pick this **only if** targeting Azure shops. |
| **[NVIDIA NCA‑GENL](https://www.nvidia.com/en-us/learn/certification/generative-ai-llm-associate/)** | $125 | medium | Cheap, recognizable brand; conceptual/associate‑level. Optional. |
| **[AWS GenAI Developer – Professional](https://aws.amazon.com/certification/certified-generative-ai-developer-professional/)** | $300 | high (AWS) | Most rigorous, but assumes ~2yr cloud + 1yr GenAI. A *later* cert, not a first. |

**Rule:** do **one** cloud/GenAI exam, matched to the cloud your target employers use. Don't collect them.

## Tier 2 — Free credentials whose **projects** are the real value

The badge is weak, but the work you produce doubles as portfolio — the best kind of credential.

- **[Hugging Face AI Agents Course](https://huggingface.co/learn/agents-course/en/unit0/introduction)** — **free cert**; the completion cert *requires a working agent*, so it's real evidence. **Do it during your agents month (Week 6).**
- **[Anthropic — Building with the Claude API](https://www.coursera.org/learn/building-with-the-claude-api)** — free on [Anthropic Academy](https://anthropic.skilljar.com/claude-with-the-anthropic-api) (your spine course); the paid Coursera cert (~$49) is optional.
- **[LangChain Academy](https://academy.langchain.com/)** — free LangGraph completion badges, if you use that framework.
- **[Generative AI with LLMs — DeepLearning.AI × AWS](https://www.coursera.org/learn/generative-ai-with-llms)** — theory (transformers, fine‑tuning, RLHF); audit free, paid cert. Fills model‑mechanics gaps.
- **[freeCodeCamp — Machine Learning with Python](https://www.freecodecamp.org/learn/machine-learning-with-python)** — free, project‑based ML fundamentals cert.
- **[Harvard CS50P](https://www.edx.org/learn/python/harvard-university-cs50-s-introduction-to-programming-with-python)** — **free Harvard certificate**, and it doubles as your **Python on‑ramp** (see `curriculum.md` Phase 0). Two‑for‑one.

## Tier 3 — Skip / don't lead with these

Fine as *free learning*; **not resume headliners — don't stack them.**
- Prompt‑engineering "get hired" badges (Vanderbilt/IBM/LinkedIn) — padding for an engineer role.
- [Google Cloud Generative AI **Leader**](https://cloud.google.com/learn/certification/generative-ai-leader) — business/strategy audience, not engineering. *(Mildly useful for your founder/pitch positioning, not an eng screen.)*
- [Azure **AI‑900**](https://learn.microsoft.com/en-us/credentials/certifications/azure-ai-fundamentals/) — starter badge; a senior can skip straight to AI‑102.
- [Kaggle micro‑course certs](https://www.kaggle.com/learn) — quick fundamentals top‑ups, negligible hiring weight.

## Recommended path for *you*

1. **Free, along the way:** the **Hugging Face Agents cert** (Week 6) — on‑path, artifact = portfolio. Plus **CS50P** (Phase 0) for a free Harvard cert while you learn Python.
2. **One paid exam, optional, near the end:** **Databricks GenAI Engineer Associate** — most aligned to your build and recruiter‑recognized.
3. **Don't over‑invest in cloud‑ML certs** unless a specific target employer requires their cloud — you're building on Claude + your own stack, not a cloud ML platform.

**Bottom line:** the shipped ReviewLoop product is your #1 credential. Certs are seasoning.