Who gets hired and what they're paid
ClearML's two director-level engineering roles, Core Platform and Web & Mobile, both list $435k–$535k (Zero G Talent's board data shows). A Staff Software Engineer tops out at $500k. The median salary across 37 salaried roles sits at $300k. A ~60-person company running an open-source MLOps platform used by 2,100+ organizations, ClearML reported, and 300,000+ AI builders, according to ClearML, pays infrastructure-tier wages because the work is infrastructure-tier: designing and operating large-scale GPU fleets, not gluing together managed services.
| Role | Location | Band |
|---|---|---|
| Director, Engineering (Web & Mobile) | New York | $435k–$535k |
| Director of Engineering (Core Platform) | New York | $435k–$535k |
| Staff Software Engineer | New York | $325k–$500k |
| Engineering Manager | New York | $275k–$350k |
| Engineering Manager, CLEAR1 (B2B) | New York | $275k–$350k |
| Enterprise Account Executive (C1 - B2B) | New York | $150k–$350k OTE |
The hiring mix tells you where the product is headed. Two engineering tracks and a thin, senior product layer. One director owns the horizontal platform — scheduler, GPU fleet manager, orchestration APIs that let customers treat a cluster like a single computer. The other owns the vertical product surfaces: CLEAR1, the B2B surface with dashboards, RBAC, and enterprise integrations distinct from the open-source core. A Solution Architecture Lead handles pre-sales and onboarding for major accounts; a DevOps Lead keeps internal CI/CD and the multi-region control plane running. Below the leads, a Backend Lead and a Group Manager signal the org has crossed the 15-to-20 threshold where flat structures fracture into pods with dedicated managers; research puts the platform-team trigger at roughly 25 engineers, the point where duplicated infrastructure work across product pods costs more than a dedicated platform group.
Non-engineering hiring is equally targeted. A VP of Sales & Partnerships and an Enterprise Account Executive with a $150k–$350k OTE band signal a land-and-expand motion aimed at Fortune 500 buyers. A VP of Finance and COO handles operations for a distributed company: 34 people in the U.S., 12 in Israel, the rest scattered across Europe, Asia, Africa, and South America. Marketing, education, and research functions appear in the department list but show no open roles, suggesting they are staffed or filled through internal mobility.
The founders set the technical bar. Moses, the CEO, brings 20 years of computer-vision and embedded-systems work, 40 patents, and an IDF elite-unit background. Gil, the co-founder, mirrors that pedigree with two decades of software R&D leadership and his own IDF service. Early hiring replicates the profile: senior ICs who can design a distributed scheduler and debug kernel-level issues, plus product managers who speak fluent PyTorch. The board's median salary of $300k confirms the market they compete in — AI infrastructure tooling, not generic SaaS.
What emerges is a hiring fingerprint: platform engineers who have built schedulers or storage systems at scale, product engineers who have shipped developer-facing consoles, and a go-to-market layer that translates GPU utilization metrics into enterprise contracts.
How the interview loop works
ClearML evaluates candidates on four pillars: technical skills, problem-solving, structured thinking, and cultural alignment. The company operates fully remotely; every stage is virtual. Candidates move through a recruiter screen, technical assessments, system design discussions, and cultural fit rounds. The exact number and depth vary by function. Data from interview preparation platforms shows the distribution of recorded experiences across roles: Engineering (14 roles), Data & ML (4), Business (4), Design (3), Product (2), and Operations (1). This spread reflects the engineering-heavy hiring mix and the premium on hands-on technical validation.
The recruiter screen is the first filter. It covers motivation for joining a fully-remote, open-source company building AI infrastructure, alignment with the mission to make infrastructure management effortless across the AI lifecycle, and a high-level review of relevant experience. For engineering roles, the conversation also probes baseline fluency in the core stack — Python, containerization (Docker, Kubernetes), cloud environments (AWS, GCP, Azure), and distributed systems fundamentals. The company explicitly requires "experience collaborating effectively in distributed or remote teams" and "excellent communication skills and ability to work independently or as part of a team" across multiple engineering postings.
Technical assessments come next, tailored to the role. Backend and platform engineers face live coding on multi-threaded or asynchronous Python, REST API design, messaging systems, and distributed architectures. Senior backend postings require 7+ years of professional Python, proven experience with MongoDB and Elasticsearch (indexing, query optimization, memory and performance tuning), and experience building and packaging Python libraries for PyPI deployment. Frontend engineers (6+ years JavaScript, 3+ years Angular, NGRX state management) work through component architecture and state management problems. DevOps and MLOps candidates demonstrate Kubernetes expertise in architecture, scaling, and operations, plus CI/CD pipeline construction, Linux administration, and database cluster management. Solutions Engineers and Architects face a hybrid: they must simplify complex technical concepts for non-technical audiences, run live demos, and translate AI/ML infrastructure capabilities into business outcomes. The careers page lists strong presentation and demo skills, plus the ability to explain such concepts to non-technical audiences, as core requirements for these roles.
System design discussions go deeper than whiteboarding. Candidates design end-to-end ML systems, justify model choices, and discuss advanced topics: RAG pipelines, ETL design for large-scale data, MLOps best practices. For ML-focused roles, the bar includes proficiency with PyTorch and Scikit-Learn, practical LLM and RAG pipeline experience, and a track record of deploying end-to-end ML systems. The company's engineering blog emphasizes that the technology developers build "quickly affects the work and productivity of people around the world" — so design reviews weigh operational realism: observability, rollback strategy, multi-tenancy, GPU utilization optimization.
Cultural fit rounds assess the "personal accountability, growth, and a culture that values diverse perspectives" highlighted on the careers page. Interviewers look for ownership — candidates who have driven projects from concept to production without hand-holding, debugged hairy distributed systems incidents, mentored peers in async environments. The fully-remote model means communication is not a soft skill; it is a survival skill. Role postings require "strong communication skills in English, both written and verbal, able to explain complex topics clearly and effectively". For customer-facing roles (Solutions Engineer, Customer Success, Sales), the bar shifts to executive-level relationship building, commercial awareness (renewals, churn mitigation, expansion), and fluency with sales methodologies like MEDDIC.
What gets candidates through the loop is consistent: demonstrated ability to ship in a remote-first, open-source environment; depth in the relevant technical domain backed by production artifacts, not just coursework; and communication clarity that scales across technical and non-technical audiences. The company hires for "curious, self-driven individuals who are excited to shape the future of AI and the infrastructure that powers it" — but curiosity without a track record of finishing things is filtered out early. Candidates who can point to specific infrastructure they built, scaled, or rescued, and explain the trade-offs they made, move fastest. Those who cannot, don't.
Where the work happens
ClearML operates a multi-hub, distributed model that mirrors the platform it builds: infrastructure-agnostic, globally accessible, designed for teams spanning on-prem clusters, cloud regions, and time zones. The physical footprint centers on three primary locations (Tel Aviv, Berkeley, Bologna) with a growing cluster in New York and a distributed workforce across major U.S. tech hubs. Each site serves a distinct strategic function, reflecting the same silicon-agnostic, cloud-agnostic philosophy ClearML sells to customers.
Tel Aviv: R&D engine
The founding headquarters sits on HaBarzel Street in Tel Aviv-Yafo, at the heart of Israel's "Silicon Wadi." Core platform engineering, research, and product strategy live here. The location is deliberate: Tel Aviv concentrates a dense pool of systems engineers, GPU-savvy researchers, and graduates from leading academic institutions who have cut their teeth on high-throughput compute and distributed systems. The HQ functions as the central hub for R&D, product strategy, engineering leadership, core software development, and global operational management. Teams designing the Infrastructure Control Plane — the layer that connects and manages GPU clusters across on-prem, cloud, and hybrid environments, sit within reach of academic labs and startup clusters pushing the same boundaries. Office culture reflects the platform's values: openness (the codebase is open-source), technical excellence, agility, robust collaboration. There is no separate research lab; the product is the lab, and engineers use it daily to orchestrate their own training runs, pipeline executions, and model-serving workloads.
Berkeley: U.S. headquarters
ClearML Inc. lists 2288 Fulton Street, Berkeley, CA 94704 as its global headquarters. This location anchors the U.S. presence: driving market penetration in the strategically important US market, providing localized expert support to American clients, building relationships within the US AI/ML community. Berkeley places the team adjacent to UC Berkeley's AI research ecosystem and the dense venture and enterprise corridor running through the Bay Area. The office serves as a base for customer-facing engineering, solutions architecture, and go-to-market leadership that translates the platform's technical capabilities, including multi-tenancy with isolated networks and storage, granular billing by compute hours and API calls, and one-click infrastructure access with built-in CI/CD integration, into enterprise deployments at companies like BlackSky, Nucleai, and Meraki.
Bologna: EMEA bridge
The EMEA office at Viale Masini 12/14, Bologna, BO 40126, Italy, extends ClearML's reach into the European enterprise and research market. Bologna's university and its growing AI and data-science community provide a talent pipeline for roles requiring fluency in EU data-sovereignty requirements, GDPR-compliant multi-tenancy, and the hybrid on-prem/cloud architectures common in European manufacturing, automotive, and public-sector accounts. The office operates as a force multiplier for the platform's "environment-agnostic" promise: engineers here validate ClearML against the sovereign-cloud stacks and regulated-data environments European customers mandate, feeding those constraints back into the core product.
New York: product and sales hub
Zero G Talent's board data shows a concentration of senior roles posted in New York: both director roles at $435k–$535k, Staff Software Engineer at $325k–$500k, both engineering manager roles at $275k–$350k, and the Enterprise Account Executive at $150k–$350k. These postings indicate a New York cluster focused on product-surface engineering (web, mobile, B2B interfaces) and high-touch enterprise sales, functions that benefit from proximity to major verticals headquartered in the city. The roles carry the same compensation bands as other hubs, signaling ClearML treats New York as a peer hub, not a satellite.
Distributed by design
Beyond the named offices, ClearML maintains a distributed presence across the United States, with personnel often located in major tech hubs working remotely. This is not a pandemic artifact; it is a structural choice matching the product. The AI Development Center is designed to be "accessible from anywhere," and the GenAI App Engine deploys LLMs onto customer clusters with ClearML handling networking, authentication, and security regardless of where the control plane runs. Engineers in major tech hubs can contribute to the core platform, the serving stack, or the Hands-On Lab (which spins up fully isolated Kubernetes clusters inside a tenant) without relocating. The distributed model also serves recruiting: the talent pool that builds large-scale MLOps tooling tends to prefer autonomy over office mandates, and ClearML's compensation, benchmarked against senior infrastructure and AI-tooling roles, reflects that reality.
The physical layout is a direct consequence of the technical problem ClearML solves: orchestrating heterogeneous compute across organizational and geographic boundaries. A team in Tel Aviv can provision a GPU cluster in a U.S. cloud region, a researcher in Bologna can launch a training job on an on-prem partition, and a solutions engineer in New York can hand a customer a one-click environment spanning both, all governed by the same control plane. The offices are not cost centers; they are testbeds. Each location runs the platform on its own hardware mix, feeding real-world friction back into the roadmap. That loop (build, use, harden, ship) is why a ~60-person company supports that many organizations and builders. The work happens wherever the GPUs are; the offices just make sure someone is close enough to the metal to keep the abstraction honest.
Who thrives here
The people who stay and advance at ClearML share a specific profile: they treat ML infrastructure as a systems discipline, not a scripting exercise. The product — a three-layer platform spanning Infrastructure Control Plane, AI Development Center, and GenAI App Engine, demands fluency across Kubernetes orchestration, GPU resource scheduling, object-storage semantics, and the full PyTorch/TensorFlow/XGBoost ecosystem. Engineers who ship here tend to have built or operated training clusters at scale, debugged distributed checkpointing failures, and designed APIs that survive version upgrades without breaking user code. The company's public commitment to backward compatibility ("all your logs, data, and pipelines will always upgrade with you") is not marketing copy; it is a constraint that shapes every design review and forces a discipline rare in MLOps tooling.
Open-source citizenship is a hiring signal, not a perk. The ClearML repository shows 111 contributors, 2,822 commits, and 194 tags across 7 branches, a cadence requiring contributors who can navigate a large codebase, write tests that catch regressions across framework versions, and review PRs for both correctness and migration safety. Staff and director-level roles are scoped for engineers who have led platform teams at companies running production GPU fleets. The $300k median compensation reflects a market that values Kubernetes operators, scheduler extensions, and Triton-backed model serving expertise at the same tier as cloud-infrastructure veterans.
Product-minded engineers thrive because the platform's five modules, Experiment Manager, MLOps/LLMOps, Data Management, Model Serving, Reports, are sold as an integrated suite. Every feature decision touches experiment tracking, pipeline orchestration, data versioning, and serving simultaneously. Candidates who have shipped developer-facing products, such as CLI tools like clearml-task and clearml-data, IDE integrations for Jupyter and PyCharm, or the orchestration dashboard that visualizes live cluster utilization, understand the feedback loop between UX and adoption. The enterprise tier adds multi-tenancy, SSO, AI Application Gateway, and Platform Management Center for tenant-level cost and usage visibility; people who have built admin planes for B2B SaaS recognize the stakes.
Customer empathy is non-negotiable. Testimonials from BlackSky (space-based imagery analytics), Nucleai (pathology AI), and Meraki (networking AI) describe workloads where GPU utilization gains and pipeline automation accelerate model delivery. Engineers who have sat on the other side of that contract, debugging stuck training jobs, explaining GPU quota overruns to finance, designing retry logic for spot-instance preemption, bring a vocabulary that accelerates design reviews. The platform's agnostic positioning (silicon, cloud, vendor, environment) means the team cannot assume NVIDIA-only, AWS-only, or Kubernetes-only; they build for the intersection, and the people who last are the ones who have been burned by lock-in before.
Retention correlates with ownership scope. The board lists those roles at $275k–$350k for both core platform and CLEAR1 tracks, signaling leads own business outcomes, not just sprint velocity. The Enterprise Account Executive band ($150k–$350k) overlaps senior engineering compensation, a structural signal that technical credibility drives sales cycles. People who optimize for narrow specialization tend to churn; those who treat "infrastructure" as everything from the driver limiting fractional GPU memory to the Markdown report a researcher shares with a stakeholder tend to stay. The company's own language ("glorious but messy process," "effortless integration," "vigorous support") selects for engineers who find satisfaction in taming complexity for others, not just publishing benchmarks.
Sixty people, five continents, two thousand customers later, the infrastructure is still disappearing — one large cluster at a time.
Working in frontier tech? Zero G Talent tracks the openings: see every open ClearML role, browse frontier tech jobs, the companies hiring, and the people building the field.