Careers at Granica: Teams, Pay and How to Get Hired

The Roles That Reveal Strategy

Granica has 11 open roles across three continents, a 45-person team, and a hiring filter that screens for one thing: verified work in the company's exact problem space. The career page splits those roles evenly across four functions — three in go-to-market, three in engineering, three in research, two in business operations — and the geography is lopsided by design. Nine sit in the Bay Area office, one in Bengaluru, one remote in the New York metro. Eight require on-site presence; two are fully remote, one hybrid. Founded in 2023, the company concentrates its senior-leaning engineering and research core in Mountain View while placing sales capacity where the buyers sit.

The filter isn't pedigree. A generalist backend engineer without Spark or lakehouse experience will not clear the keyword screen for the distributed compute role. A research candidate without large-model training runs, whether tabular or diffusion, will not reach the hiring manager. A sales applicant without enterprise AI or data infrastructure quota history will not get a first call. Granica's career page describes its teams as "small, senior-leaning" with "end-to-end ownership" and "no multi-month review cycles," meaning every hire must be productive in their specialty from week one.

The engineering roles signal the technical surface area Granica owns. Two senior software engineer positions — one focused on distributed compute and Spark systems, the other on lakehouse systems — sit at $160,000–$240,000 base, as Zero G Talent's figures put it. A forward deployed engineer role carries the same band, reflecting the expectation that engineers ship product directly into customer environments. On the research side, a research scientist working on large tabular models and another on diffusion models share that band, as does a research product manager for AI systems. Go-to-market is anchored by two enterprise account executives — one on-site in Mountain View, one remote in New York — each carrying a $280,000–$340,000 on-target earnings range, as Zero G Talent reported. A head of finance role at $140,000–$180,000, as AshbyHQ reported, and a people operations manager round out the slate.

Role descriptions spell out baseline qualifications. Research positions ask for publication records or equivalent production deployment of generative models at scale. The distributed compute role expects deep Spark internals knowledge and experience optimizing shuffle-heavy workloads. The lakehouse role requires hands-on work with Iceberg, Delta Lake, or Hudi at petabyte scale. Forward deployed engineers must demonstrate they have owned the full lifecycle of a data-intensive product in customer environments. Enterprise account executives are screened for quota attainment selling six-figure deals into data and AI leadership. The finance lead is expected to have managed fundraising and M&A modeling at a growth-stage infrastructure company.

Pay Bands: What the Numbers Show

The table combines live board postings (first six rows) with AshbyHQ listings for the final two roles, whose figures put the Forward Deployed Engineer band at $160,000–$220,000. The pattern is clear: sales carries the highest cash ceiling — $280,000 to $340,000 base before commission — while engineering and research cluster tightly at $160,000 to $240,000. The Head of Finance role sits lower at $140,000 to $180,000, reflecting a functional rather than revenue-generating track.

H1B data filed for FY 2026 corroborates the engineering median. Department of Labor LCAs for Granica Computing Inc. show a median wage of $230,000 across four job titles, with the 75th percentile also at $230,000. Those filings, while lagging real-time offers by months, confirm the company files prevailing wages consistent with the posted bands.

Equity is offered on every role listed; both board and AshbyHQ postings include "Offers Equity" as a standard line item. The benefits page describes "meaningful equity" and an employee stock purchase plan but does not publish grant sizes, vesting schedules, or strike prices. Candidates should ask for the current 409A valuation, the percentage of fully diluted shares a grant represents, and whether refresh grants are tied to performance reviews or tenure. The quarterly performance bonus applies to all roles per the benefits overview, though target percentages are not disclosed. For the Enterprise Account Executive roles, commission structure is the real variable; the posted base is only the floor.

Benefits mechanics are unusually detailed for a 45-person company. The package includes 401(k) with match, disability and life insurance, premium health, vision, and dental with FSA and dependent care, generous parental leave with a post-leave return-to-work program and onsite mother's room, fertility benefits, unlimited PTO with a four-week guideline plus quarterly company-wide recharge days, home-office stipend for remote employees, relocation assistance, daily meals (lunch and dinner) at the Mountain View office, commuter benefits, and learning stipends. The "four-week guideline" on unlimited PTO sets an expectation without a hard cap; candidates should treat it as the de facto minimum vacation they can take without signaling disengagement.

Pay transparency is listed as a formal policy, and the posted ranges are wide enough to reflect level and experience but narrow enough to prevent lowballing. A Senior Software Engineer at the top of band ($240,000) earns the same base as a Research Scientist at the top of band: the company prices systems engineering and diffusion-model research at parity. The Enterprise Account Executive band is $40,000 wider ($280,000–$340,000), which typically maps to quota attainment tiers or territory complexity. Candidates negotiating should anchor to the midpoint of the posted range for their role and ask for the compensation philosophy document; Granica's FAQ cites "fair & transparent compensation" as a stated principle, and the posted bands suggest they operate within it.

Inside the Interview Loop

Granica's interview process is intense and highly technical, reflecting the company's focus on deep research and systems engineering. Candidates move through an initial technical screen into several deep-dive rounds with senior research leadership and engineering counterparts. The pace is fast, with most candidates completing the stages within a few weeks, and the organization values people who communicate complex ideas clearly and navigate ambiguity.

The recruiting function sits with Lars Holger Steinmetzger, who said he owns every aspect of hiring from market mapping and executive search through candidate assessment, offer negotiation, and closing. That centralized ownership means the screen stays consistent across the three primary sites — Bay Area, Bengaluru, and New York — and across the evenly split mix of research, engineering, and go-to-market roles.

For research-track roles, the technical bar is explicit. The AI Research Scientist guide lists must-have skills: a PhD in Machine Learning, Statistics, or Applied Mathematics; strong grounding in information theory and statistical inference; proficiency in Python or Rust for large-scale experimentation; and hands-on experience with PyTorch, JAX, or TensorFlow. Nice-to-have signals include research experience in embeddings or model architectures for tabular data, familiarity with distributed query engines or large-scale data infrastructure, and a track record of open-source contributions or collaborative research. Deep-dive rounds map to three core assessment areas: Structured Data Modeling (demonstrating deep expertise in how relational and tabular data can be represented to enable efficient machine learning); Statistical Learning & Inference (applying learning theory to minimize resource consumption and maximize generalization); and Systems Integration (proving you can work alongside engineers to deploy your research).

Interviewers explicitly look for "Research Pragmatism." Candidates are expected to frame answers in terms of the trade-offs between theoretical performance and real-world deployment costs, and to articulate why Granica's focus on structured data is a unique and necessary evolution compared to the industry's focus on unstructured LLMs. Recent candidate reports surface questions such as "Define Model Success Metrics" and "Supervised vs Unsupervised Learning," and the topic bank tracks Structured AI / Structured Data Learning, Representation Learning, Efficient Intelligence (Resource- and Data-Efficient Learning), Symbolic + Relational + Neural Architectures (Hybrid Reasoning Systems), and Self-Optimizing Data Infrastructure.

For engineering roles posted on the Zero G Talent board, including Senior Software Engineer for Distributed Compute / Spark Systems and Senior Software Engineer for Lakehouse Systems (both Bay Area onsite at $160k–$240k), screens emphasize distributed systems fluency and data-infrastructure depth. That role (also Bay Area, $160k–$240k) adds a product-sense layer on top of the technical baseline. Go-to-market roles, including Enterprise Account Executive in New York Metro (remote) and Mountain View (onsite), both at $280k–$340k, are assessed on enterprise sales motion, technical credibility with data buyers, and ability to navigate multi-stakeholder deals.

Public signals are thin but revealing. Glassdoor shows only two interview questions and two anonymous reviews for Granica overall. TeamBlind hosts a thread with a blunt "Stay away" post from August 2024 and follow-up questions from candidates asking for detail. The scarcity of public reviews suggests either low interview volume or strong NDAs; either way, candidates should treat the published guides as the most reliable map.

What disqualifies a candidate? Absence of the PhD for research roles is a hard filter. Inability to discuss structured-data trade-offs, or defaulting to unstructured-LLM talking points, signals a mismatch with the company's pragmatic, research-driven approach. Engineers who cannot demonstrate end-to-end ownership from model to production infrastructure stall in the systems-integration round. Sales candidates without a technical sales motion in data or AI infrastructure rarely advance past the initial screen. Across functions, the company operates in a high-trust, low-bureaucracy environment; candidates who need heavy process or cannot operate with significant ownership over their work path tend to self-select out or get screened out early.

Two Coasts, One Motion

Granica operates from two offices, its Mountain View headquarters and a New York Metro presence, with a total headcount of 43 employees as of July 2026, per Tracxn. The split reflects the company's dual-engine structure: product and engineering concentrate in the Bay Area, while go-to-market roles anchor the East Coast.

The Mountain View office at 787 Castro Street houses the majority of the team. BuiltIn data shows 32 of the company's 45 listed employees sit in product and tech functions, and the job board confirms this concentration. Every technical role posted in recent months, including Senior Software Engineer for Distributed Compute and Spark Systems, Senior Software Engineer for Lakehouse Systems, Research Scientist for Diffusion Models, and the research product manager role, lists "Bay Area Office" as the location. The headquarters also hosts an Enterprise Account Executive role marked "onsite," suggesting the sales leadership or strategic accounts team keeps a foothold near the product organization. For a company building AI efficiency infrastructure that processes hundreds of petabytes of tabular data in production, per its own LinkedIn posts, proximity to the large-enterprise customers and cloud-provider ecosystems headquartered in the Bay Area is a practical advantage.

The New York Metro office operates differently. The only role currently tied to that location is Enterprise Account Executive, listed as "remote (New York Metro, remote)" with a $280,000–$340,000 band. This aligns with the June 2023 hire of an enterprise go-to-market leader as SVP of Revenue, announced via Business Wire. New York places the sales team closer to the financial-services and media enterprises that dominate the region's AI procurement budgets. The remote designation for the role also signals that Granica treats the New York site as a hub rather than a mandatory daily office, a distinction that matters for candidates weighing relocation against flexibility.

With 43 people across two sites, the company is small enough that cross-location collaboration is routine rather than exceptional. The research scientist and research product manager roles in Mountain View sit alongside the distributed-systems engineers, which suggests the LTM (Large Tabular Model) work Granica has been teasing on LinkedIn, "learning a model of the structured, tabular data universe itself," happens in close physical proximity to the systems that serve it. Meanwhile, the New York account executive feeds field intelligence back to that product loop. The geographic split maps to the two motions the company must execute simultaneously: deep technical invention on one coast, enterprise adoption on the other.

Candidates should understand that "Bay Area Office" in a posting means Mountain View, not San Francisco (a 40-minute commute without traffic), and that the New York role carries an expectation of regional travel even with the remote label. The company's own description of its Myelin agent infrastructure, launched July 2022, emphasizes durability and handoff across environments; the office structure mirrors that philosophy.

The Profile That Fits

Granica's own materials and the shape of its open roles converge on a clear profile: senior-leaning builders who want end-to-end ownership without multi-month review cycles, and who can operate across the stack from distributed systems to applied AI research. The hybrid model and remote option for sales give geographic flexibility, but the open office and daily meals in Mountain View signal that the default collaboration mode is in-person. Culture artifacts, including quarterly engagement surveys, employee awards, promote-from-history, customized development tracks, tuition reimbursement, conference access, and Lunch & Learns, are standard for a well-funded early-stage company. What distinguishes Granica is the combination of an open office floor plan, in-person all-hands, team-based strategic planning, and an explicit "open door policy" in a 45-person team. Those mechanisms only work if the median tenure and seniority are high enough that the conversation stays technical and strategic. A junior hire would spend most of their time absorbing context rather than contributing to it.

External validation is thin: a single Glassdoor review as of the latest scrape. Candidates should treat the company's own culture page as the primary source and probe for specifics in the interview: how the last strategic planning session changed the roadmap, what the last Lunch & Learn covered, whether the four-week PTO guideline is honored in practice. The compensation bands are public on the Zero G Talent board, including Enterprise Account Executive roles at $280,000–$340,000 and Senior Software Engineer and research roles at $160,000–$240,000, and the equity component is described as meaningful but not quantified in the listings.

People who thrive here tend to have a track record of owning a subsystem from design through operations, comfort arguing trade-offs with founders in real time, and a preference for density over process. If you need structure to move, this is not the place. If you move faster when the structure gets out of your way, the filter works in your favor, and the daily lunch tray at 787 Castro Street is the only onboarding document you'll get.

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