# EverAI — Tech Lead, LLM & Generative AI

| Field | Value |
|---|---|
| **Date found** | 2026-05-27 |
| **Company** | EverAI |
| **Role** | Tech Lead, LLM & Generative AI |
| **Location** | Ireland Remote |
| **Salary** | Undisclosed (B2B contract preferred) |
| **Job URL** | [LinkedIn](https://www.linkedin.com/jobs/view/4419809549/) |
| **Status** | New |

---

## Company Research

| Field | Value |
|---|---|
| **Headquarters** | Remote-first (founded by Australians) |
| **Founded** | ~2023 |
| **Employees** | ~75 (36 on LinkedIn) |
| **LinkedIn** | linkedin.com/company/everai — 22k followers |
| **Website** | — |
| **Blog** | — |

- **Product:** World's largest AI companionship / AI girlfriend platform — 50M users in 2 years, 80M+ tokens/day
- **Customers:** Consumer users seeking AI companionship (NSFW/uncensored)
- **Notable:** Top 15 fastest-growing AI companies; founding team previously exited Mad Paws (IPO), MTK Digital, Curatible (sold to Blackstone); Series unknown but fast growth

---

## Job Summary

**What they do:** EverAI runs the world's largest AI companion platform (NSFW/uncensored), processing 80M+ tokens/day across 50M users.

**The role:** ⚠️ Lead role. Tech Lead of the LLM team (3 engineers), owning model architecture, fine-tuning, RAG memory, and content moderation classifiers at massive scale.

**Core work:**
- Own fine-tuning strategy: SFT, RLHF/DPO; decide when to prompt vs fine-tune vs build new RAG pipeline
- Architect and optimise context windows, memory/RAG retrieval, and inference latency for real-time chat
- Build nuanced NSFW content moderation classifiers — context-aware, not binary filters

**Stack:** Python · PyTorch · vLLM · HuggingFace · RAG · Fine-tuning · Eval frameworks

**Work style:** Fully remote (Ireland base). B2B contract preferred. 4 weeks PTO. Annual team gathering.

---

## Score: 67%

| Dimension | Score | Justification |
|---|---|---|
| Agentic AI depth (25%) | 25% | No multi-agent orchestration; work is LLM fine-tuning, RAG for chat, and content moderation — not agentic AI |
| Tech fit (25%) | 75% | Python/PyTorch ✅; HuggingFace/fine-tuning ✅; RAG ✅; vLLM (new but learnable); no LangChain/CrewAI needed here |
| Remote fit (25%) | 100% | Fully remote |
| Company culture fit (15%) | 55% | High autonomy startup ✅; NSFW/companion AI domain is significant cultural mismatch vs Luca's compliance/enterprise background ⚠️ |
| IC/leadership balance (10%) | 75% | Player/coach: hands-on coding + 3 direct reports |
| **Final (weighted)** | **67%** | |

---

## Strengths

- Deep LLM fine-tuning (RLHF, DPO) and RAG work at extreme scale (80M tokens/day)
- Fully remote, high autonomy, fast-moving culture
- Strong HuggingFace/PyTorch stack alignment with Luca's background
- Very small team — real ownership and visibility

---

## Weaknesses & Risks

- **NSFW/AI companion domain** is a significant mismatch with Luca's enterprise/compliance AI trajectory
- Salary undisclosed; B2B contract preference adds uncertainty
- Agentic AI depth is low — pure fine-tuning + RAG, no orchestration or multi-agent work
- ⚠️ Lead role with direct reports (3 engineers)
- Company is ~75 people, very young — limited runway visibility

---

## Suggestions

- Low priority given domain and agentic depth mismatch; consider only if interested in fine-tuning specialisation
- If applying, emphasise RAG pipeline and document understanding experience; de-emphasise compliance domain
- Clarify salary expectations early (B2B rate, EUR equivalent)

---

## Interview Tracker

| Stage | Date | Notes |
|---|---|---|
| Expired | 2026-05-27 | Posting expired — no longer accepting applications |
| Applied | | |
| Recruiter screen | | |
| Technical interview | | |
| Final round | | |
| Offer / Outcome | | |