Nansen Mandates Cursor, Claude Code, and MCP Use in Senior Data Engineer Roles

Series B Funding Fuels Infrastructure Scaling Amid Institutional Demand

On December 16, 2021, Nansen closed a $75 million Series B round led by Accel, with participation from Andreessen Horowitz, Tiger Global, GIC, and SCB 10X. The Singapore-based blockchain analytics firm had secured the largest single injection of capital in its three-year history, enough to double down on a bet that institutional investors would soon demand more than price charts from on-chain data.

The funding arrived at a moment when crypto markets were fragmenting across chains, protocols, and use cases. Where Nansen had started by labeling wallets on Ethereum and Polygon, the six months between its Series A and Series B saw the company expand coverage to Binance Smart Chain, Fantom, Avalanche, Celo, and Ronin. Each new chain multiplied the complexity of its data infrastructure: transaction formats diverged, token standards evolved, and liquidity dispersed across dozens of venues. The $75 million wasn't just growth capital. It was a down payment on building systems capable of tracking 500 million labeled addresses across a multi-chain world.

Nansen's CEO Alex Svanevik framed the raise as a response to what he called "an explosion in activity in crypto, including new areas such as NFTs and DeFi." During the Series A period, the company's user base had grown over 400%, and more than 55 new team members had joined from 28 countries. But the Series B funding wasn't about keeping pace with retail curiosity. It was about preparing for institutional scrutiny. Svanevik told investors that Nansen aimed to become "investors' go-to platform for crypto insights and data, offering a trusted, single source of truth." For traditional finance firms eyeing crypto allocations, that meant moving beyond exploratory dashboards toward programmatic access through APIs and query tools.

The infrastructure implications were staggering. Nansen's data pipeline had to ingest billions of on-chain data points daily, label millions of wallet addresses, and serve real-time alerts to users tracking everything from whale movements to NFT mint patterns. The company's own post-funding roadmap included launching Nansen API and Nansen Query for institutional clients, alongside custom dashboard features that would let users function as data analysts themselves. Supporting that required not just more servers, but better architecture, systems that could scale horizontally across chains without losing the low-latency responses that institutional traders demanded.

Hiring Surge Targets Senior Data Engineers for ClickHouse, Postgres, and dbt Pipelines

Nansen's job postings make the scale explicit: the company labels and tracks 500M+ blockchain addresses, the largest such dataset in the industry, and the pipelines behind those labels move vast volumes of streaming and batch data daily. That volume shapes every technical requirement in the current hiring wave. The open Senior Data Engineer roles, based remote across Europe, ask for deep expertise in SQL and Python with hands-on production experience across ClickHouse, Postgres, and dbt, plus BigQuery in some listings. Candidates must have shipped and scaled high-performance data systems in production and demonstrate a strong grasp of streaming architectures operating at genuine scale.

The stack is not aspirational. ClickHouse handles the analytical query layer for wallet intelligence dashboards. Postgres serves transactional and metadata workloads. dbt governs transformations and modeling across the warehouse. Python glues ingestion, enrichment, and serving. The role owns data quality, reliability, and observability end to end, a point the job description underscores with a deliberate contrast: "At most analytics companies a wrong number is a support ticket. Here, it's a position someone took." That framing signals that pipeline correctness maps directly to trading risk for Nansen's institutional customers.

Hiring targets senior individual contributors with a defined runway to engineering management within 12–18 months. The company, founded in 2020 and now roughly 125 people, structures the role around full-stack data ownership: ingestion, transformation, storage, modeling, serving. Engineers partner with crypto researchers, software engineers, and product to shape how on-chain data gets modeled, surfaced, and productized. Mentorship and raising the technical bar across the team are explicit expectations.

A hard filter appears in the requirements: "You already build with tools like Cursor, Claude Code, and MCPs." Model Context Protocol integration is not optional. It is table stakes for the workflow. The team operates fully remote and async, with two no-meeting days a week protected for deep work. Most decisions happen in writing. Visa sponsorship is not provided for these roles, which narrows the candidate pool to engineers already authorized to work in Europe.

Bonus criteria reveal the domain fluency Nansen prizes: experience with on-chain data structures, Web3 technologies, and crypto markets. That combination, production-grade data engineering plus native crypto literacy, is scarce. The compensation package reflects the scarcity: competitive salary, meaningful equity, unlimited AI tokens, a personal OpenClaw instance, and a Nansen Pro account with full access to the platform's labeled dataset. Annual off-sites in Singapore, Bangkok, or Oslo (flights and accommodation covered) provide the only mandatory in-person cadence.

The hiring surge is not a volume play. Each role is built around a specific infrastructure bottleneck: scaling the labeling pipeline to absorb new chains, reducing latency on wallet intelligence queries, hardening observability so that a mislabeled address never reaches a fund's risk model. The Series B capital buys the headcount to attack those bottlenecks in parallel. The technical bar, the AI-native tooling mandate, and the async culture filter are all calibrated to the same constraint: engineers who can operate at 500M-address scale without slowing down.

AI-Native Workflow Integration: Cursor, Claude Code, and MCP as Force Multipliers

Nansen's data engineering team doesn't write SQL from scratch anymore. Instead, they prompt Claude Code in a terminal, which queries live warehouse schemas through an MCP server, generates transformation logic, and opens a pull request, all while the engineer reviews the diff. This isn't a future vision. It's the workflow Nansen's senior data engineers now use daily to ship pipelines for 500M+ labeled blockchain addresses.

The shift started in early 2026, when Cursor shipped native MCP support and the data engineering community recognized what it meant. Before MCP, AI coding assistants had no visibility into the actual data stack. Cursor could autocomplete Python and SQL, but the IDE had zero awareness of warehouse schemas, lineage, or metrics. Engineers spent hours manually syncing context between tools. With MCP, the editor sees the live state of every system the code interacts with.

Nansen built its own MCP server to expose institutional-grade blockchain intelligence directly to these workflows. The Nansen MCP server provides access to live on-chain data including transactions, token movements, wallet activities, DEX trades, and PnL across 25+ major blockchains. As Nansen's docs put it, AI tools were "flying blind to crypto markets" — they couldn't tell which wallets were accumulating or what smart money was doing in real time. The MCP server fixes that by giving agents live context.

The practical impact is measurable. Teams using Claude Code with MCP servers report pipeline debugging dropping from 2-4 hours to 15-30 minutes, dbt model authoring from 1-2 hours to 10-20 minutes, and lineage investigation from 1 hour to 5 minutes. Data quality test coverage moved from ad hoc at 40% to systematic at 90%+. These aren't marginal gains. They're the difference between iterating on a pipeline weekly versus daily.

Nansen's engineers use both Cursor and Claude Code in complementary ways. Cursor handles interactive development: writing models, exploring schemas, debugging queries. Claude Code takes over for automation: batch pipeline generation, cross-project refactors, and CI/CD integration. The shared MCP protocol means the same Data Workers configuration works identically in both tools, so context flows seamlessly between them.

This integration matters because Nansen isn't just labeling addresses. It's labeling 500M+ of them, moving vast volumes of streaming and batch data daily. The infrastructure team needs to build and optimize ClickHouse, Postgres, and dbt pipelines fast enough to keep up with institutional demand. AI agents embedded through MCP accelerate that work by turning natural language prompts into executable pipeline code grounded in real metadata.

The hiring signal is clear: Nansen's job postings explicitly require candidates who "already build with tools like Cursor, Claude Code, and MCPs." This isn't about adopting AI tools. It's about making them infrastructure. As one engineer at Anthropic noted, "It's becoming a full-time engineering discipline... teaching them how to do it is going to become the main discipline of software engineering."

For Nansen's remote Europe team, that discipline translates directly into velocity. Every prompt that generates a working pipeline, every automated lineage trace, every pull request opened by an agent is time reclaimed for deeper problems, scaling the data infrastructure behind the largest wallet label database in the industry.

Competitive Differentiation: Wallet Intelligence Over Pure Analytics

The on-chain analytics market split into three distinct product types by 2026, each serving a different job. Wallet and whale tracking platforms identify and label individual large addresses. Macro indicator platforms compute network-wide metrics like MVRV and realized cap. Raw querying tools let analysts write SQL against decoded blockchain data. Nansen competes in the first category, but it has spent years turning that category into something closer to a moat.

Nansen's edge has always been label quality. The platform attaches human-readable identities to large wallets, exchanges, market makers, funds, recognized traders, and surfaces flows in dashboards built around those labels. As of 2026, coverage spans Ethereum, Polygon, Arbitrum, Optimism, Base, and BNB Chain, with the deepest proprietary label database on Ethereum. That label density is what justifies the engineering priorities behind Nansen's $75M Series B: building pipelines that can sustain 500M+ labeled addresses means every byte of infrastructure must serve the entity, not just the transaction.

Glassnode sits in the macro category. It is the gold standard for on-chain indicators, MVRV ratio, SOPR, Realized Cap, entity-adjusted metrics, exchange flow composites, miner revenue, and holder supply segmentation. Glassnode gives fast answers with ready-made indicators, and for market-structure questions it remains unmatched. But those indicators are aggregate by design. They smooth over the individual wallet behavior that Nansen's customers, traders, funds, token analysts, actually act on. Where Glassnode answers "what stage of the market are we in," Nansen answers "who is doing something about it."

Dune Analytics occupies the raw-querying lane. It lets teams write SQL against decoded blockchain data and build custom dashboards from first principles. For protocol teams and data-heavy startups, that flexibility is essential. But it comes at a cost: adoption stalls when nobody on the team can write or validate SQL, and the time between raw blockchain activity and an actionable narrative stretches from minutes to weeks. Nansen reduces that gap by pre-labeling the entities that matter and packaging their flows into dashboards that a business team can act on immediately.

The contrast sharpens when you look at pricing and positioning. Deep Blue Alpha tracks whale wallets on Ethereum with both DEX swap and CEX flow coverage, surfacing trades through a live feed that updates continuously. Its free tier includes the live transaction feed, sentiment trends, daily whale reports, and a wallet leaderboard covering the top 50 wallets and top 25 tokens, features that CryptoQuant and Glassnode charge $29/month for. Arkham Intelligence offers the widest chain coverage with a strong free tier and a reasonable $55/month Pro upgrade. For institutional-grade wallet analytics, Nansen remains the category leader at $99–$499/month, with the same depth of proprietary labeling that defines its entire platform.

That pricing gap is not accidental. Nansen's engineering team builds ClickHouse, Postgres, and dbt pipelines not to serve generic analytics queries, but to sustain a label graph that competitors either crowdsource or approximate. Arkham uses an entity-graph approach: addresses are grouped into entities through transaction-pattern clustering plus crowd-sourced and proprietary intelligence. Arkham's intel-to-earn model has produced some of the broadest entity databases in the industry, covering 40+ blockchains. But Nansen's label quality still leads for Ethereum-focused institutional users, which is why the hiring surge targets senior data engineers who can keep that graph fresh at terabyte scale.

Most serious teams in 2026 do not pick one platform. They stack them: Glassnode for market context, Nansen for wallet intelligence, Dune for custom internal analytics. But the stack order matters. Nansen sits at the front of the workflow when the question is "who is moving capital and what does that signal?" That question drives faster trade ideas, sharper due diligence, and earlier accumulation phases. It also drives the infrastructure spend behind Nansen's Series B, pipelines built to serve entities, not just events.

The tension in the market is real: AI-assisted features have begun appearing across the category but remain early-stage at most platforms. Nansen's 2023 decision to go AI-native, before that phrase existed, put it ahead of that curve. Whether that head start translates into lasting differentiation depends on whether the label graph can scale faster than the query tools that aim to replace it.

Culture as Infrastructure: Offsites, OpenClaw, and AI Tokens as Retention Levers

Nansen does not treat remote work as a perk to negotiate around. It built a Work-From-Anywhere model from Day 1, and as of October 2023 it covered return flights plus up to one month's stay at a hub for each full-time contributor. That policy did not emerge from a pandemic retrofit. It was a compensation strategy the company chose to own outright, rather than outsourcing payroll operations to a partner while ceding control of how people are paid and where they can live.

The hubs anchor the model geographically: Singapore, Bangkok, London, and Lisbon as of 2023, with Oslo added later for annual company retreats and team off-sites. Contributors span 35+ countries, and the company runs on a universal salary scale that is location agnostic. Meritocracy rules, the policy states. That matters for European data engineers specifically because it flattens the traditional London-or-nowhere dynamic that has long distorted Web3 talent markets. A senior engineer in Lisbon can earn the same band as someone in London without relocating, and Nansen pays for the travel to bring teams together voluntarily rather than mandating presence.

The question is how you convert a distributed workforce into a cohesive one when your data teams are building pipelines behind 500M+ labeled addresses. Nansen leans on structured serendipity. It uses Donut to randomly pair folks every two weeks for coffee chats, a practice Joanna Yeoh highlighted in April 2024 as essential for fostering real connections and sparking creativity. She added that meeting colleagues in 3D sometimes just needs to happen, a direct acknowledgment that async-first culture cannot fully replace physical alignment. The counterweight is real: she also noted that coordinating those pairings across time zones was a genuine challenge, which suggests the tooling works in bursts rather than as a constant rhythm.

Offsites function as the release valve for that tension. Nansen Product Labs sets aside time every week for small teams to work on whatever they are excited about, and the annual NANC, Nansen's Awesome Premium Conference, travels somewhere together each year. These are not optional team-building exercises. They are the mechanism by which a company that has no mandatory office days maintains cohesion across terabyte-scale data flows and deep Web3 fluency requirements. People show up at the office because they want to, not because they have to.

The retention piece tightens around tooling access. As of August 2026, Nansen offered unlimited AI tokens, Claude, OpenAI, whatever helps you move fast, plus a personal OpenClaw for work and a Nansen Pro account with full access to the most detailed on-chain data in the market. That package directly matches the hiring signal from the engineering roles: candidates are expected to already build with Cursor, Claude Code, and MCPs. By giving every contributor the same AI firepower rather than rationing it, Nansen turns tool access into a retention lever that is cheaper than salary escalation and harder for competitors to replicate at scale.

The gap in the research is clear: Nansen does not provide work visa sponsorship for these roles, which constrains how deeply it can anchor European talent who are not already authorized to work in the region. That limitation sits in tension with the stated goal of attracting scarce Web3-native data talent, but it does not undermine the culture stack itself. It simply narrows the funnel.

In the end, Nansen's $75 million Series B didn't just buy headcount. It bought a workflow where the same AI agents that debug pipelines also generate them, where every senior engineer arriving from a London fintech or a Berlin startup walks into a desk with unlimited Claude tokens and a Nansen Pro account already waiting. The infrastructure isn't just ClickHouse and dbt anymore. It's the protocol that lets a prompt become a pull request, and a pull request become a label on the next wallet the market moves.

Working in frontier tech? Zero G Talent tracks the openings: see every open Cursor role, browse frontier tech jobs, the companies hiring, and the people building the field.