AI Industry Jobs
Working at companies building AI infrastructure — Anthropic, Mistral, Hugging Face, and beyond
Key insight: You do not need a PhD. ~50% of Anthropic's technical staff have no PhD. The "Inference Engineer" role barely existed 2 years ago and pays senior/staff level. Full-stack production experience is the actual bottleneck at AI companies.
$300-600K+
Total comp at frontier labs (senior level)
$100-180K
Remote from Albania at US AI companies
143.2%
AI Engineer job posting growth (2025)
$196K
Median DevRel compensation (US)
Frontier Model Companies
| Company | HQ | Remote? | Key Non-PhD Roles | Notes |
|---|---|---|---|---|
| Anthropic | SF, NYC, London | 8% remote | SWE, Platform, Trust & Safety | ~50% no PhD. 454 open roles. AI safety interest required. |
| OpenAI | SF | Mostly on-site | SWE, Platform, Residency ($220K/yr) | Residency designed for career changers. |
| Mistral AI | Paris, London, Amsterdam | Mixed | SWE, Platform | 125+ openings post-$14B valuation. EU-based. |
| Cohere | Toronto, SF, London | Remote-friendly | SWE, Platform | Enterprise-focused, more remote-friendly. |
| Google DeepMind | London, SF | Mostly on-site | Research Eng, SWE | JAX familiarity needed. <1% acceptance. |
AI Infrastructure / Inference Companies
| Company | What They Do | Key Roles | Salary Range |
|---|---|---|---|
| CoreWeave | GPU cloud infrastructure | AI/ML Infra Eng (Slurm, K8s, Ray) | $122K-$500K (median $201K) |
| Together AI | Model training & inference | SWE (Storage), Eng Manager (Model Serving) | — |
| Fireworks AI | Fast inference platform | SWE (Performance Optimization) | — |
| Modal | Serverless cloud for AI | SWE, Platform Engineer | Remote-friendly |
| RunPod | GPU cloud infrastructure | Full-stack engineers | Remote-first |
AI Developer Tools Companies
| Company | Product | Remote? | Key Roles |
|---|---|---|---|
| Hugging Face | Model hub, Transformers | EMEA remote! | Cloud ML Eng, DevRel, SWE |
| LangChain | Agent framework | Yes (open-source first) | SWE, DevRel |
| Weights & Biases | ML experiment tracking | Remote-friendly | SWE, Solutions Eng |
| Vercel | AI SDK, deployment | Remote-friendly | SWE |
| Runway | AI video generation | Some remote | AI Full Stack Eng (3+ yrs, JS/TS) |
| Comet/Opik | ML experiment tracking | Yes | DevRel Lead ($220-280K) |
Best bet for European candidates: Hugging Face explicitly lists EMEA remote positions. Mistral AI is EU-based. Both are standout options for someone in Albania.
Roles You Can Fill (No PhD Required)
Tier 1: Direct Fit (Skills Transfer Immediately)
- Software Engineer (Platform/Infrastructure) — internal tools, APIs, dashboards
- Full-Stack AI Engineer — complete apps with LLMs as core components
- Trust & Safety Engineer — abuse detection, content moderation
- Developer Experience Engineer — APIs, SDKs, documentation
Tier 2: Slight Upskilling (1-3 Months)
- AI Platform Engineer — model serving infrastructure, MLOps pipelines
- Inference Engineer — optimizing model performance, latency, cost Hot role
- ML Engineer (Applied) — fine-tuning, RAG systems, prompt engineering at scale
Salary Comparison: AI vs Traditional Tech
| Level | Traditional Tech | AI Companies | AI Premium |
|---|---|---|---|
| Junior | $80K-$120K | $100K-$150K | +25% |
| Mid-level | $120K-$170K | $165K-$220K | +28% |
| Senior | $170K-$240K | $250K-$350K+ | +40% |
| Staff/Principal | $220K-$320K | $350K-$600K+ | +60% |
Open Source Projects That Lead to Paid Work
| Project | Area | Why Contribute |
|---|---|---|
| vLLM | Inference engine | Most-used LLM serving engine. Contributors noticed by Together AI, Anyscale. |
| LangChain/LangGraph | Agent framework | Massive community, PRs noticed. Direct path to LangChain employment. |
| LlamaIndex | Data + RAG | Contributors reported job offers within 4 months of first PR. |
| Hugging Face Transformers | Model library | 100K+ stars, contributions visible to entire AI community. |
| Instructor | Structured LLM output | Small project, high impact, maintainer accessible. |
AI Safety & Alignment (SWE Roles)
What Safety SWEs Actually Do
- Build red-teaming infrastructure and automated evaluation systems
- Implement content filtering and abuse detection pipelines
- Create monitoring systems for model behavior in production
- Build interpretability tools and visualization dashboards
Key insight: Implementation roles (translating safety research into production) require senior SWE skills and are harder to fill than research roles. Getting in as a SWE first and transitioning to safety is a viable path.
How to Get Noticed
- Contribute to open source — LlamaIndex or DeepEval contributors reported job offers within 4 months
- Build portfolio projects — production RAG system, model serving pipeline, AI agent system
- Write technical content — blog posts about building AI systems in production, not tutorials
- GitHub strategy — clean code, clear READMEs, add DECISIONS.md documenting technical choices
- Network — build relationships 6-12 months before applying. AI engineering rewards connections over cold applications.
Remote-Friendliness for Eastern European Devs
| Tier | Companies | Notes |
|---|---|---|
| Fully Remote | Hugging Face (EMEA), RunPod, Modal | Best options for Albania |
| Remote-Hybrid | Together AI, W&B, LangChain, Cohere | May require US timezone overlap |
| Office-First | Anthropic (92% on-site), OpenAI, DeepMind | Mostly US/UK offices |
| EU-Based | Mistral AI (Paris/London/Amsterdam) | Best for EU work rights |