# Nebius — Senior Software Engineer (Agentic Search) - Scraping

| Field | Value |
|---|---|
| **Date found** | 2026-05-29 |
| **Company** | Nebius (Tavily team) |
| **Role** | Senior Software Engineer (Agentic Search) - Scraping |
| **Location** | European Union — Remote |
| **Salary** | Undisclosed |
| **Job URL** | https://www.linkedin.com/jobs/view/4419077091/ |
| **Status** | New |

---

## Company Research

| Field | Value |
|---|---|
| **Headquarters** | Amsterdam, Netherlands |
| **Founded** | 2023 (spun off from Yandex N.V.) |
| **Employees** | 1,500+ (400+ engineers) |
| **LinkedIn** | https://www.linkedin.com/company/nebius/ — 94k+ followers |
| **Website** | nebius.ai |
| **Blog** | nebius.ai/blog |

- **Product:** Full-stack AI cloud platform (GPU compute, storage, inference) + Tavily — search engine for AI agents providing real-time web access optimised for LLMs and RAG
- **Customers:** AI developers, enterprises, and researchers building and deploying GenAI applications at scale
- **Notable:** Listed on Nasdaq (NBIS), HQ Amsterdam; $2B NVIDIA investment; Tavily team builds the web access layer for AI agents; backed by leading investors

---

## Job Summary

**What they do:** Nebius's Tavily team builds the search engine for AI agents — the web access layer that gives LLMs and agentic systems real-time, structured knowledge of the internet.

**The role:** Senior IC on the Scraping team within Tavily, owning the distributed web acquisition pipelines that power AI agent search at scale.

**Core work:**
- Build distributed data acquisition systems to capture and structure the live web for AI agent consumption
- Investigate and circumvent browser fingerprinting, anti-automation systems, and evolving web protections using Playwright/Puppeteer/CDP
- Design and deploy high-performance, resilient scraping orchestration layers with stealthy, adaptive behaviour

**Stack:** Python · Node.js · Playwright · Puppeteer · Chrome DevTools Protocol · Kubernetes · Docker

**Work style:** Fully remote within the EU; fast-moving, IC-owned systems with direct impact on AI agent capabilities

---

## Score: 65%

| Dimension | Score | Justification |
|---|---|---|
| Agentic AI depth (25%) | 35% | Enables AI agents by providing web access, but the actual work is browser automation and web scraping — infrastructure one layer below agent engineering |
| Tech fit (25%) | 45% | Python/Node.js overlap; browser automation is specialised and not in the target stack (no LangGraph/LangChain/Claude/PyTorch) |
| Remote fit (25%) | 100% | Fully remote EU — no location friction |
| Company culture fit (15%) | 72% | Nebius/Tavily is fast-moving AI-native; Tavily is a startup within Nebius with high ownership culture |
| IC/leadership balance (10%) | 90% | Pure IC builder role, full ownership of scraping systems |
| **Final (weighted)** | **65%** | |

---

## Strengths

- Nebius/Tavily is genuinely AI-native — the search engine for AI agents is a critical piece of the agentic AI stack
- Fully remote EU with no restrictions
- High autonomy and ownership culture; small team within a well-funded company (Nasdaq-listed, NVIDIA-backed)

---

## Weaknesses & Risks

- Core daily work is browser automation and web scraping — peripheral to agentic AI engineering (building the infrastructure that agents use, not building agents)
- No LangGraph/LangChain/Claude stack in this role; target tech stack is browser/network automation
- Role requires deep specialisation in browser internals and anti-automation systems — a different skill domain from AI engineering
- Salary undisclosed

---

## Suggestions

- Only apply if interested in pivoting toward search/retrieval infrastructure for AI agents
- Emphasise distributed systems experience and any web scraping / browser automation background
- Ask about the Tavily product roadmap and where the scraping team sits relative to the agentic/LLM features

---

## Interview Tracker

| Stage | Date | Notes |
|---|---|---|
| Applied | | |
| Recruiter screen | | |
| Technical interview | | |
| Final round | | |
| Offer / Outcome | | |