Who gets hired
Mariana Minerals closed a $310 million Series B in August 2026 and immediately posted 39 salaried roles across five U.S. facilities, with a compensation band of $63,000 to $211,000 (median $170,000) per Zero G Talent's board data, for engineers, geologists, and operations specialists to advance mineral extraction and processing technologies at copper, lithium, nickel, cobalt, aluminum, graphite, uranium, and rare earths projects. This piece maps the hiring picture: who gets hired, what it pays, the six-round interview process, where the work happens at five sites, and why the work-life score lags behind mission and culture ratings.
The mining industry has watched its talent pool shrink for decades. The a16z interview transcript notes the labor pool is contracting across both trades and engineering skill sets, and that mining has not been a magnet for talent. Mariana entered that vacuum with a different proposition: a minerals company that hires like a technology company, because it operates like one.
Mariana calls itself the world's only software-first minerals project developer and operator. The San Francisco headquarters houses strategy, software, and metallurgy. The Houston office runs piping, stress analysis, process modeling, and project controls. Three project sites (Lithium One, Copper One, and a third stealth asset) host pilot plants and field operations. As of August 2026, the company lists 177 employees, according to LinkedIn, on LinkedIn; Zero G Talent's board tracks the 39 salaried roles.
The role taxonomy reads like a cross between a mine development roster and a modern engineering org chart. Recent postings span Sr. Technical Sourcer and Warehouse & Logistics Coordinator in San Francisco; Process Safety Engineer, Engineering Manager for Piping & Stress Analysis, and Staff Process Modeling Engineer in Houston; plus Materials Engineer, Sr. Electrical Engineer, Metallurgical Engineer, and Pilot Plant Technician (weekend shift) at the San Francisco facility. A Lead Resource Geologist search runs separately through Ashby for open-pit and underground technical studies. First-party board data shows senior software and program roles, including Head of Engineering Programs ($180k, according to Zero G Talent's board–$240k, Zero G Talent's board figures put), Staff Machine Learning Engineer in Ann Arbor ($160k, Zero G Talent's board data shows–$240k), and Senior Full Stack Software Engineer ($150k–$220k, Zero G Talent's board reported), confirming the software-first label extends beyond marketing.
The founder's background explains the blend. After a decade at Tesla across cell manufacturing, cathode production, and refining, he observed that mining lacked the vertical integration and automation culture that battery manufacturing forced into existence. "The degree to which automation was absent was pretty interesting," he said in a 2025 interview. That gap shapes hiring: the company needs geologists and geophysicists, mining and geotechnical engineers, process and chemical engineers, chemists and metallurgists, mechanical, structural, and civil engineers, plus the long tail of technicians and operators who move material from point A to point B. But it also needs machine learning engineers who can apply reinforcement learning to plant control, full-stack engineers who can replace disaggregated data systems with unified software, and LLMs that eliminate manual data entry between databases.
Team boundaries reflect that duality. The Houston office functions as a traditional engineering house (piping, stress, process modeling, project controls), while San Francisco concentrates metallurgy, pilot operations, and the software layer that wraps both. The project sites demand field-ready generalists who can commission equipment, troubleshoot circuits, and feed data back to the models. A Pilot Plant Technician on weekend shift in San Francisco earns $40–$50 an hour; that Houston role commands $100k–$160k. The spread is deliberate: the company pays for physical competence and software fluency in the same budget.
What holds the org together is a stated mission to supply minerals critical to modern energy, AI, and defense technologies, and a hiring philosophy that treats the labor shortage as a design constraint. That constraint drives the compensation structure.
What it pays
Mariana Minerals' compensation structure reflects a capital-intensive hard-tech operation that needs senior engineering talent willing to work on-site at industrial facilities. Zero G Talent's board data shows a salaried band running from $63,000 to $211,000 with a median of $170,000 across 39 roles. That median sits above the $140,000–$175,000, NewJob reported, posted range as of July 2026, suggesting the public postings skew toward mid-level individual-contributor positions while the board also captures higher-band leadership and specialized roles.
| Role | Location | Salary Band (USD/year) | Source |
|---|---|---|---|
| Head of Engineering Programs | San Francisco HQ | $180,000 – $240,000 | Zero G Talent board |
| Staff Machine Learning Engineer | Ann Arbor, MI | $160,000 – $240,000 | Zero G Talent board |
| Senior Full Stack Software Engineer | San Francisco HQ | $150,000 – $220,000 | Zero G Talent board |
| Project Controls Manager | Houston, TX | $171,000 – $216,000 | Zero G Talent board |
| Sr. Electrical Engineer | San Francisco HQ | $140,000 – $210,000 | Zero G Talent board |
| Engineering Manager, Piping & Stress Analysis | Houston, TX | $170,000 – $210,000 | Zero G Talent board |
The spread reveals two things. First, geography doesn't compress the band — Houston roles sit at the same ceiling as San Francisco HQ roles. Second, the top of the range ($240,000) appears on both a software-focused ML role and a hardware-focused engineering-programs role, signaling that Mariana values cross-domain leadership equally. Hourly postings on LinkedIn (San Francisco $32–$42/hr, $40–$50/hr) translate to roughly $66,000–$104,000 annualized, consistent with the board's $63,000 floor for salaried roles and likely representing technician or operator classifications.
Third-party data aligns directionally but lags in granularity. AskTheRecruiter's July 31 snapshot of 31 open roles reported a median posted band of $140,000–$175,000; newjob.tech shows the same median with a top-decile marker at $240,000+. The board's $170,000 median across 39 salaried roles is higher, which makes sense: the board ingests the company's full requisition set, including unadvertised or recently filled positions that aggregators miss. SimplyHired lists 83 "Mariana minerals jobs" but without salary granularity; LinkedIn's recent postings show San Francisco bands at $100,000–$140,000 and $140,000–$175,000, plus a Houston band at $100,000–$160,000, all sitting inside the board's wider envelope.
A $310 million Series B closed in August 2026, atop a $73.5 million private placement that April, gives the company runway to sustain these bands. In a sector where incumbents such as Freeport, Rio Tinto, and BHP outsource engineering to consultancies and startups compete for the same shrinking pool of mining-engineering talent, Mariana's compensation sits at a premium to traditional mining salaries but below top-tier Bay Area software packages. That positioning is deliberate: the company needs engineers who can operate in "dirty spaces" — plants, yards, remote sites — and the pay reflects the scarcity of people willing to do that work at software-grade technical depth. The six-round interview process tests for exactly that willingness.
Six rounds, six weeks
Mariana Minerals runs a six-round process that typically spans four to six weeks, per interview guides updated August 2026. The sequence is deliberate: an initial qualification screen, technical evaluations, behavioral assessments, a scenario-based problem-solving session, collaborative meetings with cross-functional peers, and a final decision round. Dataford aggregates eight interview reports and rates overall difficulty at 5.4 out of 10; Glassdoor's nine reviews show a 44 percent positive experience rating and a 2.67-out-of-5 difficulty score. Candidates for machine learning and process engineering roles consistently report the toughest interviews, while engineering interns and senior software engineers describe a more moderate experience.
The first screen filters for baseline qualifications and alignment with the company's mission to reimagine the minerals supply chain through software, automation, and data-driven decision-making. Technical evaluations then diverge by discipline. Data engineers face questions on pipeline architecture, ETL/ELT patterns, and system design. Machine learning candidates are tested on model deployment, feature engineering, and production monitoring. Project managers confront critical path analysis, earned value management, and complex resource leveling across multiple site locations. Full-stack engineers work through frontend architecture, API design, and database scaling scenarios. Across roles, the topic distribution leans heavily toward system design (100 percent of ML and data engineering loops), project controls (100 percent for PMs), and coding fundamentals (95 percent overall).
Behavioral assessments follow a structured rubric. The company explicitly prioritizes candidates who treat project management as a "disciplined, repeatable science" and who can provide quantifiable evidence of past successes. Interviewers ask for specific examples: establishing project baselines and handling deviations, managing budgets exceeding $10 million, balancing speed against technical quality, communicating delays to senior leadership, and fostering accountability without direct authority. Safety compliance is non-negotiable — any answer that omits safety protocols is flagged as a significant red flag. The STAR method (Situation, Task, Action, Result) is recommended for structuring these responses.
Scenario-based problem solving tests how candidates operate under the constraints of Mariana's integrated work culture. You might be asked to audit a failing project and name the first three analysis points, pivot strategy due to regulatory change, or resolve conflict between engineering and operations teams. Collaborative meetings then put you in front of representatives from both sides (engineers, procurement specialists, field operations managers) to evaluate whether you can lead through influence when timelines compress.
Preparation guidance from the company's own interview guides is unusually specific: dedicate at least two weeks, master your chosen methodology (Waterfall or Agile) and defend its fit for the project type, bring three to four detailed stories covering budget overruns, schedule slips, and team conflicts, and prepare insightful questions about the current project pipeline or technology integration roadmap. Quantification is expected — "reduced project cycle time by 15 percent through optimized resource scheduling" beats "improved efficiency." The process concludes with standard pre-employment screening for a safety-sensitive, drug-free workplace. The work itself happens across five sites built for this integration.
Where the work happens
Mariana Minerals operates across five primary sites (two corporate offices and three project locations), each serving a distinct function in the company's vertically integrated model. The split reflects a deliberate structure: software and capital-project engineering concentrate in the coastal offices, while the physical extraction and refining infrastructure sits in the interior West and Gulf Coast.
San Francisco headquarters anchors the software and autonomy stack. The roles posted there (Head of Engineering Programs, Senior Full Stack Software Engineer, Staff Machine Learning Engineer, Sr. Electrical Engineer) signal the concentration of MarianaOS development: MineOS for autonomous fleet and mine-planning integration, PlantOS with its reinforcement-learning process-control layer, and CapitalProjectOS for LLM-driven engineering-procurement-construction workflows. The office functions as the control-room brain for the field sites; a feature shipped here can be observed moving recovery rates or reagent consumption at Copper One within days.
Houston engineering office houses the capital-project execution muscle. Postings for Project Controls Manager and Engineering Manager, Piping & Stress Analysis confirm the focus: translating the digital models built in San Francisco into procurable, constructible, commissionable hardware. Houston's proximity to the Gulf Coast industrial base and the company's lithium pilot work in the region makes it the natural home for the piping, structural, and project-controls disciplines that turn a process flowsheet into a built plant.
Copper One, near Moab, Utah, is the only fully operating mine-refinery complex in the portfolio today. Acquired in late 2025 after 15-plus years of prior operation, it runs an open-pit mine feeding a heap bioleach circuit and a hydrometallurgical refinery using solvent extraction and electrowinning to produce high-purity copper cathode for the U.S. market. The site is the live testbed for MineOS (autonomous fleet, integrated geologic models, real-time mine planning) and PlantOS (dynamic process control, equipment-health monitoring). In mid-2026 the company began deploying Boston Dynamics' Spot quadruped for inspections — the first visible step toward its stated goal of an end-to-end autonomous mine and refinery.
Lithium One is the first greenfield commercial facility, targeting first production in the first half of 2027. The process extracts battery-grade lithium salts from oil-and-gas wastewater — a chemistry the company has already validated in lab and pilot units. PlantOS's reinforcement-learning platform will manage the variable feed compositions, shifting ore grades, and fluctuating flow rates that make conventional human-controlled operations unstable. The facility is designed from the ground up as a digital-twin-native asset: every sensor, valve, and pump feeds the model that will eventually run the plant with no human in the loop.
Stealth 3 remains undisclosed in public materials, but its placement alongside Copper One and Lithium One on the company's project list indicates a third mine or refinery in active development. Given the stated roadmap (covering the full suite from copper to rare earths) and the goal of ten projects in ten years, Stealth 3 is likely the next commodity-specific facility entering the CapitalProjectOS pipeline. The Houston office will manage its engineering-procurement-construction execution; San Francisco will deliver the autonomy stack; the field site will generate the operational data that closes the flywheel for the next project.
The geographic separation is not administrative convenience. It is the physical manifestation of a software-first mining company: the code lives where talent density is highest; the metal is made where the ore and energy are. That manifestation shows up in the work-life numbers.
Why the work-life score lags
Glassdoor data tells a clear story: 79 percent of Mariana employees would recommend the company to a friend, and culture and career opportunity both sit at 3.8 out of 5. Work-life balance trails at 3.4. People who stay and advance at Mariana tend to treat the mission as a genuine filter, not a slogan.
The company describes itself as a software-first, vertically integrated minerals business supplying critical inputs for that mission. That phrasing is deliberate. "Software-first" in a sector that has historically run on spreadsheets and site visits means the engineers and scientists who thrive here are comfortable writing code that talks to furnaces, not just dashboards. The board data bears this out: the open roles cluster at senior and staff levels (Staff Machine Learning Engineer at $160k–$240k, Head of Engineering Programs at $180k–$240k, Senior Full Stack Engineer at $150k–$220k) and they sit across San Francisco, Ann Arbor, and Houston. A candidate who needs hand-holding on the hardware side will stall; the ones who move fast are the ones who have already broken a model against a physical process and lived to debug it.
Vertical integration changes the collaboration model. A piping stress analysis manager in Houston ($170k–$210k) and a project controls manager in the same office ($171k–$216k) are not handing off specs to a contractor — they are iterating with the same metallurgists and ML engineers who sit in San Francisco. The 2025 acquisition of the Lisbon Valley copper operation in Utah, now running as Copper One, added a live cathode-producing asset to the portfolio. That means the geologists and process engineers who join are not modeling hypothetical ore bodies; they are optimizing a running plant. The people who last are the ones who treat a plant trip as a data set, not a crisis.
Mission alignment shows up in the hiring loop. The company's public materials repeat the same phrase: "supply the critical minerals powering those sectors." Candidates who can map their last project directly to that sentence clear the behavioral screen faster.
The 3.4 work-life balance score is not an accident. Five U.S. facilities, a recent acquisition integration, and a tech stack that spans mine-to-market mean the pace is set by physics and customer delivery dates, not sprint calendars. The engineers who thrive have already made peace with that trade-off — who chose the constraint because the problem is harder and the impact is measurable in tons of cathode copper, not click-through rates.
Working in frontier tech? Zero G Talent tracks the openings: see every open Mariana Minerals role, browse frontier tech jobs, the companies hiring, and the people building the field.