Working at dili: Culture, Pace and Who Thrives

How Work Gets Done

The job board tells a story the website doesn't: six distinct functions, seven open roles, a salary band stretching from $83,000 to $222,000. For a company that lists two founders on its team page and shows 61 employees on LinkedIn, that spread reflects a team operating in a frontier compliance-AI domain where a single error risks billions in tax credits.

dili was founded in 2023. Zero G Talent's first-party board data shows active postings across demand generation marketing, DevOps, senior full-stack engineering, senior QA, product management, and compliance customer success, all based in New York or remote within the U.S. The median posted salary sits at $190,000 (Zero G Talent's figures put the median at $190,000). That functional breadth across a headcount of 15–61 means each role carries significant scope.

The coordination structure is implicit in the roles themselves. Zero G Talent's board data shows no engineering manager, no engineering lead, no VP of product. Remote work is baked into the postings: every role lists "New York, NY, US / Remote (US)" as location. The hiring bar, visible in the "senior" prefixes and the salary bands, selects for people who have already operated autonomously in their function. The board's median of $190,000 reflects a market rate for that autonomy in New York.

What the postings don't show is how the founders allocate their own time across product, sales, fundraising, and operational demands. But the functional spread of the open roles makes the boundary clear: the founders own the problems that don't fit into the six defined seats. Everything else ships through the person sitting in that seat.

Values That Keep the Math Working

Dili's operating principles read like a case study in constraint-driven clarity. The company's public positioning (website copy, founder interviews, investor updates) converges on a handful of values that function as survival mechanisms for a team processing 700+ projects across a regulated energy-infrastructure market.

The foundational value is domain depth over model breadth. "We automated what everyone said couldn't be automated, in an industry where the rules were being written in real time," the founders write on their about page, referencing their prior work building compliance systems at Coinbase. That experience (automating more than $2 trillion in compliance at a global financial exchange) shapes every architectural choice. When TechCrunch asked CEO Anand Chaturvedi about AI's role in July 2026, he drew a sharp line: contemporary models handle only the data layer, translating unstructured documents into structured data. From there, a deterministic system applies the "complex-but-static compliance rules." The distinction matters. In prevailing wage and apprenticeship compliance, a hallucinated wage rate doesn't just produce a wrong answer — it puts tax credits at risk. The team's choice to confine generative AI to ingestion, not judgment, is a value statement: reliability trumps novelty.

Customer proximity as a design constraint appears repeatedly. The platform's "five minutes a week per contractor" metric isn't marketing fluff; it's a survival requirement. The drag-and-drop payroll upload, the direct SAM.gov integration for real-time wage rates across all 50 states, the support for WH-347 and state-specific forms — each feature exists because compliance software fails when the people submitting data resist it. "Our subs actually use it. That's the difference," a CFO at a top-10 national EPC firm told the company. A project manager at an electrical contractor put it more bluntly: "Every Friday I spend five minutes uploading our reports and I'm done. My phone used to ring all weekend with compliance questions. Now nobody's calling me about issues anymore."

Data privacy as non-negotiable emerged early, when the product still targeted private equity due diligence. In a February 2024 TechCrunch piece, co-founder Stephanie Song addressed the sector's skepticism head-on: private customer data isn't used to train Dili's models, many of which are open source, and the company plans to offer funds the ability to create their own models trained on proprietary, offline data. That commitment carried into the infrastructure pivot. When a Bloomberg Law survey found 30 percent of deal lawyers wouldn't consider AI at any stage of due diligence over confidentiality concerns, Dili's response was architectural. The deterministic compliance engine means customer data never touches the probabilistic layer for rule application.

Flexible commercial models reflect a team that meets customers where they are. Chaturvedi told TechCrunch in July 2026 that roughly half the projects use Dili as an in-house tool; the other half outsource the entire compliance function to Dili on a contractor basis. The company supports both. "The interesting thing will be how the market itself evolves and where the customer needs go as AI develops," he said. The value isn't model purity — it's meeting the customer's organizational reality, whether that's a compliance team wanting tenfold capacity (the website claims five-minute average review versus seven-plus hours manual) or a developer needing full-service monitoring to protect tax credits.

Measurable outcomes over activity metrics shows up in the numbers the company chooses to publish: 94.7 percent overall compliance score, 97 percent prevailing wage compliance, 89 percent apprenticeship compliance, 48x faster review, 158 percent ROI on fines alone. These aren't vanity metrics; they're the language of the buyers (EPCs, developers, tax equity investors) who live or die by audit results. The team processes thousands of payroll reports covering tens of thousands of workers. The specificity is the point. In a regulated industry, vague claims get discarded; precise, auditable claims get purchased.

Hiring from the regulated industry is the least visible but most consequential value. The founding team didn't just recruit engineers; they brought in former Department of Labor enforcement officials, EPC compliance leads, apprenticeship specialists. The website states it plainly: "We combined compliance automation expertise with construction industry veterans." That blend means the product encodes regulatory logic that pure software teams would miss: the nuance of RAPIDS certification validation, apprentice hour percentage tracking, ratio compliance monitoring, state program integration. It also means the sales conversation starts from shared vocabulary, not translation.

Together these principles form a coherent operating system for a team in a massive, regulated market. They don't call them values. They call them the only way the math works.

The Hiring Bar

The salary bands detailed above tell the first story. Seven salaried roles spanning $83,000 to $222,000 with a $190,000 median signal a team that hires senior individual contributors who can operate without scaffolding. In a company spread across six functional areas (engineering, product, sales, customer success, marketing, compliance), there is no one to hand work to. The hiring bar targets those who have built the thing, sold the thing, or supported the thing at scale elsewhere.

The second signal comes from the problem space. dili automates due diligence for VC and private equity funds, an industry sitting on $311 billion in unspent cash while deploying the lowest capital total in seven years ($67 billion), TechCrunch reported in February 2024. Gartner predicts three-quarters of VC executive reviews will be AI-informed by 2025. Yet 30 percent of deal lawyers still refuse AI at any diligence stage, citing confidentiality risks. dili's pilot ran with 400 analysts across funds and banks. The company fine-tunes open-source models to cut hallucination and keeps customer data out of training runs. Candidates who get through understand that accuracy isn't a metric here — it's the product. A hallucinated clause in a term sheet isn't a bug; it's a lawsuit. The bar rewards engineers who have shipped in regulated or high-liability domains, and non-technical hires who can translate between fund partners and model outputs without losing either audience.

The third signal is structural. The only public employee review available on Indeed for a company named "Dili" (dated July 2022) describes "strict quality unachievable quality targets" and "moving goalposts during probation period to avoid giving pay increases." That review appears to reference a different entity (a business in Liv, not the AI compliance company founded in 2023) and reflects one experience. The research provides no verified internal employee accounts for dili-ai.

The fourth signal is mission alignment. Customer testimonials on dili's site ("We had two years of payroll data that nobody had reviewed. Dili ran a look-back assessment and surfaced issues in four days that would have taken our consultants months") reveal the actual work: compressing months of specialist labor into days of automated review. The people who thrive are the ones who find that compression energizing, not terrifying.

The fifth signal is team fluency. In a company of this scale, "team" means the whole company. There are no silos to hide in. A DevOps engineer who can't explain a deployment to the Compliance Customer Success Manager creates a bottleneck the company can't route around. The bar selects for people who have operated in genuine cross-functional settings — not "I collaborated with design once" but "I shipped a feature that required negotiating scope with sales, compliance review with legal, and instrumentation with SRE, and I owned the timeline."

None of this is written on a careers page. It's legible in the salary bands, the pilot scale, the regulatory constraints, the public testimonials, and the company's own descriptions of how work gets done. The hiring bar doesn't select for potential in the abstract. It selects for demonstrated ability to deliver senior-level output across functional boundaries while the target moves — because that is the job.

What the Reviews Don't Tell You

The public record on dili's employee experience is fragmented across two Glassdoor entities and a LinkedIn presence that tells a different growth story than the job board's hiring signal. As of the latest data, Glassdoor lists four anonymous reviews for "Dili Development Company" and 36 for "DiliTrust"; neither of which maps cleanly to the "dili-ai" LinkedIn page showing 61 employees and a $15 million Series A closed in August 2026. The review counts alone suggest two things: the DiliTrust entity has accumulated enough feedback for pattern recognition, while the Development Company entry is too thin to generalize from. Neither Glassdoor page is dated in the research digest, so the vintage of those reviews is unknown; they could predate the 2023 founding date on LinkedIn or reflect a different corporate structure entirely.

What the first-party board data does show, concretely, is the salary band detailed above, competitive for New York–based frontier-tech roles and implying a company budgeting for experienced hires, not junior generalists. The breadth of functions (engineering, product, compliance, marketing) also contradicts any "small team spread across six functional areas" framing if the LinkedIn headcount reflects current full-time employees rather than cumulative hires, contractors, or a related entity. A core team of 15–61 does carry a DevOps specialist, a QA lead, and a dedicated compliance success manager simultaneously, which the board confirms.

The LinkedIn growth claims sharpen the tension. The company's own August 2026 post states "500% growth in 6 months" to over 700 supported projects, with hiring active across engineering, customer success, and business development. That trajectory (from a 2023 founding to 61 LinkedIn profiles and a $15 million Series A closed in August 2026) describes a venture-scale operation. If the Glassdoor reviews for DiliTrust (36 reviews) belong to this entity, they capture a workforce larger than that signal. If they belong to a separate DiliTrust product line or predecessor, the signal is noise.

No verbatim employee quotes appear in the research. The TechCrunch 2024 profile quotes founder Stephanie Song on analyst pain points ("Analysts burn the midnight oil working hundreds of hours doing the work that nobody wants to do") but that is a market observation, not an internal testimonial. The LinkedIn posts from 2026 highlight contractor satisfaction ("Contractors consistently prefer working within Dili's platform") and technical throughput ("18 months of certified payroll processed in under 24 hours"), again external-facing.

The gap is real: the research provides review counts, salary bands, and growth claims, but no attributable employee voice (positive or negative) tied to a name, role, or date. Until Glassdoor review text or dated employee interviews surface for dili-ai specifically, the two-sided picture this section aims to deliver cannot be grounded. The most honest read is that the public record, as of August 2026, documents a rapidly scaling compliance-AI company with competitive pay and an ambiguous review footprint — and a job board that still shows seven open roles across six functions for a company with 15–61 employees.

Who Stays, Who Leaves

The profile that succeeds at dili looks less like a traditional specialist and more like a compressed generalist — someone who can move between a deterministic rules engine, a customer success escalation, and a product decision without asking for a handoff. The company's CTO, Brian Fernandez, has described the stack explicitly: "customer success, compliance expertise, and engineering." That triad isn't organizational chart decoration; it's the daily reality for a team that grew 500% in six months after the Series A led by Khosla Ventures and now supports over 700 energy and infrastructure projects representing more than $1 billion in gross wages reviewed.

People who thrive here tend to arrive with one of two backgrounds — or the rare ability to simulate both. The first is direct domain fluency: contractors, payroll administrators, DOL auditors, CPAs who have lived the "certified payroll in every format imaginable" problem that dili's AI normalizes before its deterministic engine flags violations. The job posting for a Compliance Customer Success Manager explicitly targets candidates who have "spent your career working with contractors and chasing certified payroll reports." The second background is high-velocity engineering (DevOps, senior full-stack, senior QA) but with a tolerance for the unglamorous work of data ingestion pipelines that turn messy subcontractor timesheets into auditable records. Product managers sit in the middle, translating between the compliance team's "final judgement call" and the engineering team's retrieval pipelines.

What burns people out is the refusal of the work to stay in its lane. A transformer repair in year three of a solar facility's operation can put a 30 percent tax credit at risk, dropping it to 6 percent — a liability that surfaces years after construction closeout when "compliance oversight typically diminishes." The compliance experts who make the final judgement call depend on the deterministic engine producing "the same result for the same inputs every time," traceable to a specific rule. There is no throw-it-over-the-wall moment. The 500 percent growth rate means the wall keeps moving.

The hiring signals reflect this. The Demand Gen Marketer role exists because the market (49 gigawatts of utility-scale solar, wind, and storage installed in 2025, with a threefold pipeline projected for 2026–2027) is moving faster than inbound can handle. But the marketer can't just run campaigns; they need to understand why a GC or EPC on a clean energy build carries liability for "every sub's certified payroll, every wage determination, every labor hour." The same holds for the DevOps engineer who inherits infrastructure that must be "consistent and auditable" because backpay and penalties depend on the outcome.

Founders Stephanie Song, Brian Fernandez, and Anand Chaturvedi built this from Coinbase corporate development, where Song watched analysts handling the grunt work.

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