Lakebase Has Thousands of Customers and Its Own Job Category. Neither Existed Before October.

What the $134 Billion Signal Means for Hiring

Databricks closed 2025 by raising $4 billion in a Series L round at a $134 billion valuation. TechCrunch reported the round valued the company at $134 billion, up 34% from the $100 billion valuation it had achieved just three months earlier. The company's revenue run rate crossed $4.8 billion in Q3, growing more than 55% year over year. For context, Snowflake, Databricks' closest public competitor, reported $4.7 billion in full-year fiscal 2026 revenue and trades at a market cap roughly half of Databricks' private valuation.

The company has raised over $7 billion in combined equity and debt, including a $1.8 billion debt facility led by JPMorgan in January 2026 that bankers read as a pre-IPO signal. CEO Ali Ghodsi told CNBC he "wouldn't rule out a 2026 IPO" and confirmed the company is "IPO-ready" with board structure, auditing, and financial reporting in place. The Wall Street Journal reported Databricks plans to use the new capital to add thousands of jobs across Asia, Europe, and Latin America, and to bring on more AI researchers.

That hiring ambition is already visible. Zero G Talent's board shows 43 Databricks roles added in the past week alone, spanning senior forward-deployed engineering positions across US financial services hubs (Atlanta, Boston, Chicago, Dallas, Philadelphia) with salaries listed between $180,656 and $248,360. These aren't speculative postings. They map directly to the company's stated expansion priorities: financial services, enterprise AI deployment, and geographic reach beyond the Bay Area.

The financial profile behind the hiring push is unusual for a company at this stage. Databricks is free-cash-flow positive on a trailing 12-month basis while growing at 55%+. Its data warehousing product crossed $1 billion in run rate in under four years from general availability. Lakebase, its serverless Postgres database for AI agent workloads, has thousands of customers just six months after launch and is growing revenue at twice the pace of the warehousing product.

Databricks is not recruiting to sustain a business. It is recruiting to staff a land grab. Ghodsi himself used that phrase in the CNBC interview — "It's kind of a land grab, with do-it-yourself winning right now" — referring to the wave of enterprise developers building AI applications on their own data. The $4 billion in fresh capital, the pre-IPO debt facility, and the accelerating revenue curve give Databricks the runway to hire aggressively across engineering, security, and forward-deployed roles before a public listing resets the compensation calculus.

The question for infrastructure talent isn't whether Databricks is hiring. It's whether the roles being created now, in AI-native database engineering, security lakehouse architecture, and enterprise agent deployment, represent the leading edge of what the post-hype AI stack actually requires.

Lakebase: A Six-Month-Old Product That Created a New Job Category

Databricks didn't just build a new product with Lakebase. It created a new job category that didn't exist six months ago and now sits at the center of the company's most aggressive hiring push.

Lakebase is Databricks' managed PostgreSQL transactional layer, built to run OLTP workloads directly inside the lakehouse. Instead of running a separate operational database and piping data into the lakehouse for analytics and AI, Lakebase lets applications, analytics, and agents operate on the same fresh data in one place. The October 2025 acquisition of Mooncake Labs, a San Francisco startup focused on making Postgres data immediately usable across lakehouse workloads, accelerated that timeline and deepened the technical surface area Databricks now needs to staff against.

On Databricks' own careers board, roles with "Lakebase" in the title now span solutions architecture, sales specialization, technical solutions engineering, and staff-level database engineering. The Solutions Architect - Lakebase listing, based in London, requires 10 years of transactional database expertise (OLTP engineering, product development, administration, and pre-sales) plus experience integrating transactional systems into big-data, lakehouse, and AI ecosystems. It's not a generalist data-engineering role. It's a specialist position built around a product that merges two worlds most enterprises have kept separate.

The Staff Database Engineer role goes deeper on the core platform side: advising software engineering teams on functional and non-functional feature designs, validating that designs satisfy customer requirements, and building operational tools for troubleshooting and enterprise extensions. This is the kind of role that exists because Lakebase isn't a thin wrapper. It's a managed PostgreSQL service with its own performance, availability, and migration challenges.

Then there's the support layer. The Senior Technical Solutions Engineer, Lakebase position, posted in Mountain View, is explicitly a customer-facing technical support role for Lakebase and the Neon cloud-native Postgres service it runs on. Candidates need strong SQL database experience, Linux administration skills, hands-on cloud platform work, and the ability to troubleshoot complex architectural scenarios in real time. It's a role that exists at the intersection of database internals and customer escalation, a profile that's been common at companies like AWS and Snowflake for years but is new territory for Databricks' hiring pipeline.

What makes Lakebase hiring distinct from the rest of Databricks' recruitment surge is the specificity of the domain knowledge. The company's broader data-engineering roles ask for distributed systems experience, Spark, Delta Lake, and cloud infrastructure. Lakebase roles add a hard requirement for transactional database internals: disaster recovery, high availability, backup and recovery, scale-out methods, identity and security management, and vendor-to-vendor or on-prem-to-cloud migrations. These are skills that live in the operational-database world, not the analytics world, and they signal that Databricks is staffing for a product that has to earn trust with the engineers who run production OLTP systems.

The Mooncake Labs acquisition added another dimension. Mooncake's technology is designed to make that same cross-environment data flow seamless, which means Lakebase engineers now need to think about how transactional data flows into analytics and AI pipelines without the latency and complexity of traditional ETL. That's a systems-integration problem that pulls in streaming, data modeling, and agent-framework knowledge, broadening the hiring profile beyond pure database engineering.

For infrastructure engineers watching the market, Lakebase roles represent something rare: a genuinely new product category creating genuinely new job profiles, not just a rebrand of existing data-engineering work. The question is whether the category sticks and whether the engineers hired to build and support it are building the foundation of how enterprises run transactional AI workloads, or staffing a product that gets absorbed back into the broader lakehouse platform within two years.

Panther Labs: When Security Operations Became a Data Engineering Problem

Databricks' June 16 agreement to acquire Panther Labs marks the company's third security acquisition in under a year, and it signals something specific about where enterprise AI infrastructure is heading: security operations are becoming a data engineering problem, and the talent to solve it is suddenly scarce.

Panther, valued at roughly $1.4 billion after a $120 million Series B in 2021, built an AI SOC platform around detection-as-code and cloud-native telemetry. Its platform runs inside customers' own AWS accounts against their Snowflake or Databricks warehouses, meaning security data never leaves the customer's environment. That architecture made it a natural fit for Anthropic, which uses Panther to manage security operations for its frontier AI infrastructure. Tim Nguyen, Anthropic's Head of Defense, said Panther brought "a software engineering approach to detection and response" that let his team adapt as the environment evolved.

The acquisition accelerates a product roadmap Databricks already started. In March, the company launched Lakewatch, its security lakehouse designed to unify security, IT, and business data into a single governed environment for agentic detection and response. Panther adds the SOC layer on top: automated alert triage, context gathering, and response recommendations, all running as AI agents against the data Lakewatch centralizes.

Ghodsi framed the rationale bluntly at the Data + AI Summit in San Francisco. "If they're going to attack you with agents, you have to defend with agents," he told Reuters. "You have to fight fire with fire."

The deal puts Databricks in direct competition with CrowdStrike and Cisco-owned Splunk, both of which still dominate the legacy SIEM market. But the underlying shift is bigger than any single vendor. Legacy SIEMs force security teams to limit what data they ingest because of cost and complexity. That creates blind spots across cloud services, SaaS applications, identity providers, and AI systems, exactly the surfaces attackers are now probing with AI agents.

For hiring, the implications are concrete. Panther's team includes engineers and former SOC analysts with deep experience in open source and cloud-native security operations. Jack Naglieri, Panther's founder and CEO, previously led StreamAlert, an open source project originally developed at Airbnb. Those engineers now fold into Databricks' security organization, which is building out a platform that needs people who understand both security telemetry at scale and the lakehouse architecture it runs on.

The skill set is narrow. Databricks' security lakehouse work demands engineers who can build detection-as-code pipelines, integrate telemetry from cloud infrastructure and identity providers, and design agentic workflows that automate SOC operations. It is not traditional security operations work. It is infrastructure engineering with a security domain, and the talent pool is thin.

The Panther acquisition suggests the volume of open roles will keep climbing, and the positions will increasingly blend security operations expertise with data platform engineering, a combination the market has not yet learned to produce at scale.

Agent Bricks and the Enterprise-Agent Sales Engineering Pipeline

Databricks launched Agent Bricks at that same June 2025 event, and the product is already reshaping who the company hires and what those people do. The platform, a no-code/low-code framework for building, deploying, and governing AI agents on enterprise data, sits at the center of a new class of roles that didn't exist six months ago.

The pitch is straightforward. Agent Bricks uses a company's own schemas, business definitions, and custom semantics to help agents pick the right tools, join tables correctly, and return answers that are grounded in actual enterprise context rather than generic model output. It supports models from OpenAI, Anthropic, Google, Meta, and open-source providers through a single contract, with Unity Catalog enforcing permissions and rate limits per user or team. Governance (prompt injection prevention, sensitive data detection, content filtering) runs through the same control plane that governs the data itself.

That integration is what makes the hiring different from a typical AI platform play. Agent Bricks doesn't just need ML engineers who can fine-tune models. It needs people who understand enterprise data architecture well enough to wire agents into it, and sales engineers who can explain to a Fortune 500 CTO why an agent built on their own lakehouse won't hallucinate its way through a compliance audit.

The customer case studies Databricks publishes make the scope clear. AstraZeneca used Agent Bricks to parse over 400,000 clinical trial documents and extract structured data points without writing code, in under 60 minutes, said Joseph Roemer, Head of Data and AI, Commercial IT. Hawaiian Electric built a "Regulatory Chat Tool" against 40,000 legal documents that outperformed their previous open-source setup on accuracy and setup time, said Joel Wasson, Manager of Enterprise Data and Analytics. YipitData replaced regex-based ETL pipelines with agentic reasoning and got 20x more coverage with better accuracy, said Edward Goo, Head of Data Engineering.

Each of those deployments required someone who could translate a business problem into an agent workflow, and that's the role Databricks is now hiring for at scale. The LinkedIn job board shows parallel demand: Delivery Solutions Architects in Stockholm, Software Engineer roles for GenAI inference in San Francisco, and Systems PhD-level engineers in Seattle.

The "Forward Deployed Engineer" title is telling. It's a role borrowed from the Palantir playbook: an engineer embedded with the customer, building and iterating on production systems in real time. At Databricks, these roles sit at the intersection of Professional Services, Engineering, and Sales. They need to understand Lakebase for agent memory, Mosaic AI Vector Search for retrieval, Unity Catalog for governance, and the Agent Bricks framework itself, then explain all of it to a financial services client who cares about audit trails more than benchmarks.

On the sales side, the company is building out specialized roles around its new product lines. LinkedIn lists a Director of Lakebase Sales Specialists for the Central region, based in Munich, and a Lakebase Sales Specialist in Zurich, roles that didn't exist before the product launched. Agent Bricks creates the same demand curve: someone has to sell the platform, which means someone has to understand it deeply enough to run a proof of concept on a prospect's actual data.

The technical bar for these roles is specific. Agent Bricks supports the Model Context Protocol (MCP) for tool integration, connects to APIs and SaaS applications through Unity Catalog-managed credentials, and traces every agent interaction in MLflow. Engineers working with it need to be comfortable with LangChain, LangGraph, or LlamaIndex for orchestration, and with Databricks' own stack: Delta Live Tables for data prep, Mosaic AI Vector Search for semantic retrieval, and Databricks Apps for building user-facing interfaces on top of agents.

Prakash Trivedi, writing on LinkedIn about building agentic systems on Databricks, described the full pipeline: ingest raw data through Delta Live Tables, vectorize it with embedding models, design agent logic in notebooks, orchestrate multi-step reasoning with LangGraph or Agent Bricks, evaluate with the Agent Evaluation Framework, trace everything in MLflow, and deploy through Mosaic AI Gateway as REST endpoints. That pipeline is essentially the job description for the roles Databricks is filling now.

The company's own documentation frames Agent Bricks as the thing that turns weeks of manual work into minutes. Lippert's Director of AI, Chris Nishnick, said the platform let them productionize domain-specific agents for tasks like extracting insights from customer support calls, work that used to take weeks of manual review. Franklin Templeton's Lead Data Scientist Colin Zimmerman said fund analysis that took days now takes seconds, with every insight grounded in their data and business logic.

That compression (weeks to minutes, days to seconds) is what enterprises are buying. And it's what's driving Databricks to hire the people who can deliver it. The Agent Bricks launch didn't just add a product to the lineup. It created an entire go-to-market and delivery motion around enterprise-agent engineering, and the hiring reflects it: forward deployed engineers, solutions architects, sales specialists, and GenAI inference engineers, all converging on the same problem of making autonomous agents work inside the constraints of real business infrastructure.

Where the Jobs Actually Are: Financial Services and EMEA

Databricks' hiring blitz isn't spread evenly. Two patterns stand out: financial services roles are the single largest vertical concentration, and the EMEA region, anchored by a massive UK investment, is where the company is adding headcount fastest outside the US.

Start with financial services. The most telling titles are the cluster of "Sr. Forward Deployed Engineer - Financial Services" positions, posted for Atlanta, Dallas, Boston, Chicago, Philadelphia, and a central US role, all with salary bands between $180,656 and $248,360. That's not experimental hiring. That's a sales-engineering army being built to sell into banks, insurers, and asset managers. LinkedIn's job listings confirm the pattern: an EMEA Financial Services Leader role based in London, plus a dedicated "Databricks for Financial Services" showcase page that didn't exist two years ago. The company's own careers site names financial services accessibility as one of the core problems its platform solves, alongside drug discovery and agriculture, a signal that the vertical is now central to the pitch, not an afterthought.

The EMEA story is bigger in scale and more deliberate in execution. In March 2026, Databricks announced an $850 million investment in the UK over three years, quadrupling its London office footprint and designating the city as its EMEA headquarters. The London office, opened in 2024 at nearly 30,000 square feet, is now the hub for a region that spans Amsterdam, Madrid, Milan, Belgrade, Munich, and a new engineering facility in Zagreb. Dael Williamson, Databricks' EMEA CTO since 2023, has overseen the buildout. The LinkedIn job board reflects it: Lakebase Sales Specialists in Zurich, a Director of Lakebase Sales in Munich, a Delivery Solutions Architect in Stockholm, roles that map directly onto the office expansion.

The Middle East is quieter but present. Bayt.com lists over 1,020 Databricks roles across the Middle East and Gulf, though that figure likely includes partner and consulting positions alongside direct hires. What's clear is that the MEA region is being built out as a sales and solutions-architect coverage zone, not an engineering center; the technical roles still cluster in Bengaluru, London, and the US.

The takeaway: if you're an infrastructure engineer or sales engineer targeting Databricks, financial services expertise and EMEA availability are the two signals that will get your resume noticed. The company is hiring generalists too, but the concentration is unmistakable.

What the Job Postings Actually Require

Strip away the hype-cycle language on Databricks' careers page ("innovators, builders and truthseekers") and the actual job postings tell a plainer story. The company is hiring for a stack that barely existed three years ago, and the specific technical demands reveal what production-grade enterprise AI looks like when the proof-of-concept phase ends.

Start with Lakebase. The product is a serverless Postgres-compatible database that handles OLTP workloads inside the Databricks lakehouse. Building and supporting it means hiring engineers who understand both transactional database internals and distributed lakehouse architecture, not one or the other. The Lakebase SKILL.md on GitHub reads like a job description in disguise: candidates need to manage compute units scaling from 0.5 to 112 CU, configure copy-on-write branching for CI/CD environments, handle OAuth token refresh for Postgres connections, and debug schema permission conflicts between service principals and human developers. PostgreSQL 16 or 17, psycopg 3.x, the Databricks CLI at v0.294.0 minimum, and the Python SDK at 0.81.0: the version floors alone tell you how fast the platform is moving.

Then there are the integrations. Lakebase isn't a standalone database. It syncs Unity Catalog Delta tables into Postgres for low-latency app reads, captures Postgres changes back into Delta via change data feed, serves as an online feature store for ML models, and stores state for AI agents built with LangGraph or the OpenAI Agents SDK. Engineers working on this surface area need to understand Delta Lake, Unity Catalog governance, change data capture pipelines, and vector similarity search via pgvector, all in one role. The Microsoft documentation for Lakebase on Azure Databricks lists these as separate integration tracks. In practice, a single team owns them all.

The Panther acquisition adds another layer. Security lakehouse work demands people who can operate across log analytics, SIEM-style detection rules, and data infrastructure, translating raw telemetry into governed, queryable tables inside Unity Catalog. Databricks' careers page lists Security as a distinct department alongside Engineering and Field Engineering, which signals the function has moved from a support role to a core product pillar.

On the customer-facing side, the most revealing titles are the Forward Deployed Engineer roles. These aren't sales engineers who demo a product and leave. They embed with clients to architect solutions on the Databricks platform, which means they need enough depth in data engineering, ML ops, and security to build production pipelines under real enterprise constraints. The range of city postings (Atlanta, Boston, Chicago, Dallas, Philadelphia) maps directly to major financial services hubs, confirming where the hardest enterprise adoption problems live.

The broader skills gap is well-documented. RevolentGroup has written about the Databricks skills shortage creating fierce competition in hiring. McKinsey's estimate, cited by Databricks itself, says as many as 375 million workers globally may need to change occupations to meet evolving company needs. That number is abstract, but the version floors on the Lakebase CLI and SDK are concrete: if your Postgres knowledge stops at version 14 and your Python SDK experience is two releases behind, you're already out of date.

What emerges is a profile that didn't have a name five years ago: infrastructure engineers who bridge OLTP and OLAP, who understand both the database layer and the AI agent layer, and who can operate inside a governance framework (Unity Catalog, HIPAA, C5, TISAX) without treating it as someone else's problem. Databricks isn't hiring generalists. It's hiring people who can hold the full stack in their head, from compute unit sizing to agent state management, and ship under enterprise deadlines.

The Bigger Picture: An Industry Crossing a Threshold

Databricks isn't hiring in a vacuum. The company's recruitment surge mirrors a structural shift happening across the entire enterprise AI stack. The industry has crossed a threshold: 78% of organizations now run AI in production environments, up from 55% the year before, according to Stanford HAI's 2025 AI Index Report. But that top-line number hides a brutal reality.

Only 5% of custom enterprise AI tools ever reach production deployment. That figure, from MIT's State of AI in Business 2025 report and corroborated by Cleanlab's survey of 1,837 engineering leaders, defines what everyone in the space now calls the "GenAI Divide." Companies can build demos. They can spin up pilots. What they can't do (what almost none of them can do) is turn those prototypes into systems that reliably execute authenticated actions across Gmail, Slack, Salesforce, and the 106 other SaaS apps the average enterprise runs. This is the gap Databricks is hiring to fill, and it's the same gap that makes its $134 billion valuation defensible.

The skills demanded by Databricks' open roles tell you where the bottleneck actually is. These aren't prompt engineers or fine-tuning specialists. They're infrastructure engineers who understand OAuth lifecycle management, data pipeline architecture, and the security requirements of regulated industries. The Cleanlab survey found that 70% of regulated enterprises rebuild their AI agent stack every three months or faster, a churn rate that makes permanent, production-grade infrastructure nearly impossible without dedicated talent. Fewer than one in three teams report satisfaction with their observability and guardrail solutions. Reliability, not model capability, is the weakest layer in the stack.

This is why the Panther acquisition matters beyond the security use case. Panther brought Databricks a team that had already solved the problem of deploying and monitoring data infrastructure at scale in production, exactly the competency that separates the 5% from the 95%. And it's why Lakebase is generating its own hiring track: the ability to run transactional and analytical workloads on a single platform removes one of the most persistent integration failures that keeps custom tools stuck in pilot purgatory.

The broader market data reinforces the pattern. U.S. private AI investment hit $109.1 billion in 2024, with generative AI attracting $33.9 billion globally, an 18.7% increase from 2023. Yet 95% of organizations report that current AI spending has produced little to no measurable business return. The money is flowing in. The value is not flowing out. The constraint isn't capital or model quality. It's the infrastructure and security engineering talent required to deploy systems that work when real users and real data are involved.

Index.dev's adoption data puts the trust problem in stark relief: only 27% of organizations say they trust fully autonomous AI agents, down from 43% a year earlier. Trust is declining even as capability improves, because the failures that matter (data leaks, hallucinated outputs in customer-facing workflows, broken authentication chains) become visible only in production. Databricks' hiring concentration in financial services, where regulatory scrutiny makes those failures expensive, is a direct response to this dynamic.

The companies that will capture value from the next phase of enterprise AI aren't the ones with the best models. They're the ones that can deploy, monitor, and govern AI systems in production with the same rigor they apply to any other critical infrastructure. Databricks is betting that this infrastructure layer, and the engineers who build it, is where the next $100 billion in enterprise AI value will be won or lost.

Working in AI? Zero G Talent tracks the openings: browse AI jobs, openings at Databricks, and the people building the field.