Mach9’s resume screen rewards one thing most applicants overlook

Nine Roles, Three Tracks

Mach9 is hiring for nine roles right now, eight of them on-site in San Francisco, and its screening funnel is built to filter out anyone who hasn't shipped perception systems that survive contact with real infrastructure data. The company turns raw reality-capture data into engineering-ready 3D maps at 96× the speed of manual drafting, and its product sits where computer vision meets civil infrastructure and production-grade CAD. That technical surface area dictates who they need, and why the roles on its careers page don't resemble a generic startup's "we're growing" spreadsheet.

The open positions cluster into three tracks: commercial, engineering, and operations. On the commercial side, five roles span the full go-to-market motion. An Account Executive and a Solutions Engineer (both San Francisco, on-site) will work the direct sales cycle. A Solutions Consultant (remote) supports technical validation for prospects. A Community Marketing Manager (San Francisco, on-site) owns the practitioner-facing brand. A Head of Growth (San Francisco, on-site) sits above the funnel, accountable for pipeline architecture rather than individual quota.

Engineering carries the heaviest load. The Head of ML Engineering (San Francisco) leads the perception stack that extracts 3D object and line features, including breaklines, curbs, striping, poles, signs, and wires, from dense LiDAR point clouds and imagery. A Machine Learning Engineer (San Francisco) builds and trains those 3D scene-understanding models directly, leveraging what the company calls a "unique data advantage" from 100,000-plus miles surveyed and 11 million-plus features extracted across 2,500-plus projects. A QA Survey Technician (San Francisco) validates extraction quality against the same hardware partners (Riegl, Trimble, Leica) that customers use in the field. A Software Engineer in Test Automation (SDET) (San Francisco, on-site) builds the regression harness that keeps the CAD pipeline stable as model versions ship.

Operations rounds out the slate with two roles that signal organizational maturity. A Founding Recruiting Lead (San Francisco) will design the hiring engine that feeds the technical teams above. A People Operations Lead (San Francisco, on-site) handles the internal systems that keep a 96× speed culture from burning out the people who sustain it. Two additional listings, a Contract Software Engineer for Internal Tools (remote) and a Lead Product Designer (San Francisco, on-site), appear on the board but sit outside the core nine-role count referenced in the company's public messaging.

Every role is full-time. Equity packages extend to all full-time employees. Relocation assistance is offered for San Francisco-based positions. The concentration of on-site requirements reflects a team that still ships hardware-adjacent software from a single lab, not a distributed collective.

How the Funnel Works

Mach9 does not publish a public playbook for its hiring funnel. Glassdoor lists three interview questions and three anonymous reviews for Mach9 Robotics, confirming a multi-step process exists, but the specific questions and exact stage gates are not disclosed. The careers site describes the work as "frontier perception challenges" with "real ownership and fast iteration," and notes that "the pace is fast, but the bar for quality is even higher." Those two signals — speed and quality — are the only documented criteria candidates can reasonably expect to face at every stage.

The initial resume screen filters for the hard technical prerequisites listed across the nine open roles: perception, sensor fusion, 3D reconstruction, geospatial data pipelines, and production-grade C++ or Python. The careers page emphasizes that the platform "accelerates the creation of engineering deliverables from raw data, cutting manual drafting time by 96×" and that customers include "global leaders in engineering and construction" with "professional liability on every deliverable they ship." That context means the screen looks for evidence of shipping code that touches physical-world data (LiDAR, photogrammetry, point clouds) not just model accuracy on benchmarks. A candidate whose background is purely research-grade perception without deployment experience will likely stall.

Beyond the resume, the interview loop tests two things in parallel: technical depth on the specific stack Mach9 uses (Riegl, Trimble, Leica hardware; large-scale 3D mapping; 11 million features from those projects) and the ability to operate at the advertised pace. The careers page repeats its core value — speed, customer-driven iteration, and an uncompromising quality bar — and "close isn't good enough." That phrasing signals interviewers will probe for stories where the candidate delivered a working system under tight deadlines, handled noisy real-world data, and made trade-offs without sacrificing correctness. The employee testimonial that "you'll thrive if you want to work with strong people and ship meaningful systems end-to-end" reinforces that the loop evaluates ownership: can you take a problem from raw sensor input to a deliverable an engineer stamps?

Glassdoor's three reviews are too few to reconstruct a reliable stage map, but their presence implies at least a phone screen, a technical deep-dive, and a final round, typical for a Y Combinator-backed startup of this size. The careers page's emphasis on "the toughest challenges demand the brightest minds" and "we invest in exceptional people and let the best ideas lead" suggests the final round includes a conversation with a founder or senior technical lead, where evaluation shifts from pure technical execution to judgment: how you scope ambiguous problems, how you prioritize when everything is urgent, and whether you can defend a technical decision to a customer who carries professional liability.

What is not in the public record, and what this article will not speculate on, are the exact number of stages, pass rates at each gate, specific coding challenges, or whether a take-home project is required. The research simply does not support those details. What the research does support is a funnel that mirrors the product: high-throughput, precision-oriented, and unforgiving of "close enough." Candidates who advance will be those who can show, not tell, that they have built and shipped such systems.

Signals That Clear the Bar

Mach9's screening funnel filters for evidence that a candidate can operate at the intersection of production machine learning, geospatial data at infrastructure scale, and the zero-tolerance quality bar that civil engineering demands. The company's own descriptions — "frontier perception challenges," "real ownership and fast iteration," "never compromise on excellence because the work demands it" — map directly to the signals reviewers look for in the first pass.

Production ML that ships to paying engineers

The Head of ML and ML Infrastructure Engineer roles make the priority explicit: Mach9 builds perception models that extract 3D object and line features from such data, then serves those models to surveyors and engineers who carry such liability. Candidates who advance past the initial screen typically show a track record of taking models from experiment to production, not just publishing papers. The research page notes the company's "unique data advantage" and emphasizes that the work is "both research-driven and" production-facing. A resume that lists only academic benchmarks or Kaggle rankings stalls; one that describes deploying point-cloud segmentation models to customers who measure success in centimeters and liability exposure moves forward.

Geospatial fluency, not just CV chops

Customer testimonials repeatedly cite the domain gap: "resolving edge-case geometry," "statewide inventories," "design-grade deliverable." The screen rewards candidates who understand the difference between a neat point-cloud demo and a corridor project a DOT will stamp. Experience with civil infrastructure workflows (transportation and DOTs, telecom, utilities, survey & engineering) appears in every customer quote and in the "Built for the infrastructure work your team delivers" positioning. Familiarity with the sensor ecosystem (Riegl, Trimble, Leica) or with reality-capture pipelines (mobile mapping, aerial LiDAR, photogrammetry) is a concrete signal. So is any background in the software tools those firms actually use: Civil 3D, MicroStation, TerraSolid, or the data-exchange formats (LandXML, IFC, LAS/LAZ) that move between field and office.

Speed-with-quality discipline

"Our company, like our software, is built for speed. We work fast, iterate with customer feedback, and never compromise on excellence because the work demands it." That value statement appears verbatim on the careers page and is echoed by employees: "This sentiment is echoed." The initial screen looks for proof you have operated under that tension. Shipped a model that cut a month-long manual process to days? Documented the validation protocol that kept false positives below the liability threshold? Led a sprint where the definition of done included a QA review by a licensed engineer? Those specifics survive the screen; "fast learner" and "high standards" do not.

Customer-proximity instinct

Multiple employees describe the loop: "listen closely to people who use our technology and turn their challenges into ground-changing solutions," "customers leave a meeting excited to use a data extraction software again," "directly review Mach9's AI-generated assets." The Solutions Engineer role exists because the product cannot succeed as a black box. Candidates who have sat across from a civil engineer, a survey crew chief, or a GIS specialist and translated their pain into a spec, or who have built tooling those users adopted without a training program, carry a strong signal. The screen treats that exposure as a proxy for the "human approach" the company claims.

End-to-end ownership in a small, senior team

"Mach9 is a home for builders who want ambitious problems and the freedom to solve them… real ownership and fast iteration… ship such systems." The current headcount — nine open roles across engineering, product, GTM — implies a team where every hire expands the surface area they own. The screen favors candidates who have owned a feature from problem definition through deployment and post-launch monitoring, especially in a sub-50-person startup. Y Combinator backing and the investor roster (Cruise, Autodesk, Adobe, DoorDash founders) signal network density; the screen checks whether you can operate in that density without hand-holding.

What the screen does not weight

The research contains no mention of specific degree requirements, years-of-experience minimums, or algorithmic puzzle performance. The company's public materials lead with outcomes, such as miles surveyed, projects delivered, and weeks shaved, and with the hardness of the problem ("Teaching machines to read the physical world is hard… Close isn't good enough"). The initial screen is calibrated to those outcomes. A candidate who can show, with numbers and named customers or projects, that they have moved a geospatial ML system from prototype to liability-grade production clears the bar. Everything else is noise.

Building an Application That Lands

Mach9's careers page states the bar plainly: "This is hard, and our customers carry such liability. That's not enough." That sentence should shape every line of your application. The company builds AI-powered geospatial tools that turn lidar and imagery into engineering-grade CAD and GIS deliverables — cutting drafting time 96×, per their own figures — and serves firms like Langan, Olsson, Woolpert, HDR, and multiple state DOTs. Their product, Digital Surveyor, extracts 11 million-plus features across 2,500-plus projects and 100,000-plus surveyed miles, running on hardware from Riegl, Trimble, and Leica. The team traces its roots to Carnegie Mellon's Robotics Institute and a Y Combinator batch, backed by Quiet Capital and operators from Cruise, Autodesk, Adobe, and DoorDash. Your resume, portfolio, and cover letter need to show you can operate at that intersection of perception, infrastructure, and production-grade software.

Resume: map experience to the stack

Lead with the specific domains Mach9 works in. If you've processed lidar point clouds, built feature-extraction pipelines, or shipped cloud-based 3D applications, put those projects in the first third of the resume. Name the sensors (Riegl, Trimble, Leica, or equivalent), the data formats (LAS/LAZ, E57, proprietary), and the downstream consumers (Civil 3D, MicroStation, ArcGIS, QGIS). Quantify scale: miles of corridor, number of features, latency targets, accuracy thresholds. The company's own metrics, centimeter-level accuracy at infrastructure scale and 10× faster extraction per technician, are the language they speak. If you've reduced manual drafting hours, automated QA workflows, or integrated with CAD APIs, state the before/after numbers.

Engineering roles should surface the hard constraints you've worked under: real-time inference on large point clouds, GPU memory budgets, coordinate-system fidelity, multi-sensor fusion. Product and GTM candidates should highlight experience selling to or supporting civil engineering firms, survey shops, or DOTs, the buyer universe Mach9 lives in. The seed round of roughly $2.5 million closed in December was earmarked for "new hiring across both technical and business development roles," per Technical.ly, so both sides of the house are active.

Portfolio: show the pipeline, not the demo

A GitHub repo with a clean README beats a polished video. Include the data ingest path, the model or heuristic that extracts linear features (kerbs, pavement edges, guardrails, barriers), and the export format a civil engineer can open and stamp. Mach9's Digital Surveyor 2 adds an AI-powered extraction engine that "suggests linear features directly from Lidar point clouds, adapting in real time to each individual user's drawing patterns," according to GIM International. If your portfolio demonstrates adaptive learning from user corrections, or a human-in-the-loop verification loop that matches their "verify vertices and lines in 3D" workflow, call it out explicitly. Russell Hall of Langan called Digital Surveyor 2 "potentially the lowest-barrier-to-entry software in its category"; so if you've built tooling that lets field surveyors with limited CAD training become productive fast, that's a direct signal.

Cover letter: write to the problem, not the brand

Skip the generic admiration. Address the bottleneck Mach9 exists to solve: "Conventional surveying is only as fast as you can hire. Reality capture moved the bottleneck to the office." That framing comes from their own site. Connect your background to one of their three stated product purposes (improving access to geospatial content, delivering insights from maps and models, or integrating with established solutions) as Baikovitz described to Geo Week News. If you've worked inside a state DOT, a geospatial firm, or a civil design shop, name the deliverable you produced and the pain point you lived. The team quotes on the careers page reinforce what they value: "real ownership and fast iteration" (Praveen Venkatesh), "quality results, delivered quickly" (company value), "carry such liability" (Baikovitz). Mirror that vocabulary when you describe your own work.

Format and signals

Keep the resume to one page if you're under eight years out; two if you've led teams or shipped multiple products. Use a clean, machine-readable layout with no columns that break ATS parsers. Link your portfolio and GitHub in the header. If you've published at CVPR, ICRA, or a geospatial venue (ASPRS, ISPRS), list it. If you hold a PE, LS, or GISP, include it; Mach9's client base respects those credentials. The company's own employees (Seth Gulich, Max Leung, Brandon Cheng, Shayne Shen, Alex Fischer) emphasize "hard problems," "freedom to solve them," and "exceptional people do their best work." Your application should read like someone who has already chosen the hard problem and shipped the solution.

What This Guide Does Not Cover

This article has one job: explain what gets a candidate past Mach9's initial screen and through the funnel. It does not cover compensation benchmarks, benefits packages, work-life balance, or whether the ping-pong table sees regular use. Those topics have their own literature and their own data sources, but they do not change the screening criteria.

The salary data that exists is scattered across third-party aggregators and varies by entity. The table below summarizes the cited figures; these figures are real, they are cited, and they are irrelevant to whether your resume clears the first filter.

Mach9 Robotics shows only three reported salaries on Glassdoor.

Benefits and perks appear on Mach9's own careers page (health, vision, and dental coverage, paid parental leave, flexible time off) and are summarized on Built In, Wellfound, and Glassdoor. Culture reviews live on those same platforms. None of them appear in the screening rubric the hiring team uses when they open a PDF at 9 a.m. on a Tuesday.

The article also does not address: equity structure, vesting schedules, refresh grants, 401(k) match, office location policies, remote-work eligibility, team offsites, learning stipends, or the founder's leadership philosophy. Candidates who need that information before applying should consult the sources above. Candidates who want to pass the screen should study the signals in the sections above, and then prove they've already shipped what Mach9 needs.

Working in AI? Zero G Talent tracks the openings: see every open Databricks role, browse AI jobs, openings at Anthropic and Harvey AI, and the people building the field.