Questions we get asked, answered without the hedging

Redcubical Systems Private Limited designs, builds and operates custom software, cloud platforms, data systems and applied AI for client organisations, and runs 4 commercial software products of its own. Eleven service lines cover build, run and intelligence work. The products matter commercially because they mean we make the same architecture, cost and reliability decisions on our own balance sheet every week.

Our head office is at Dev Heights, First Floor, Aman Vihar, Sahastradhara Road, Dehradun, Uttarakhand 248001, India, with a second engineering hub in Bengaluru, Karnataka. All delivery happens from those two locations. We serve clients in 10 countries and hold committed daily overlap with each market rather than operating overseas offices we would have to staff thinly.

Email nitesh@redcubical.com or call +91 90687 14658 with a short brief. The sequence is a 45-minute technical call with the architect who would lead the work, then a written proposal within five business days carrying a fixed price or a capped ceiling, a named team roster and a delivery calendar. Nothing up to that point is charged.

Within one business day for any enquiry, and usually the same day if it arrives before 15:00 IST. A technical call is normally scheduled inside four working days. Security questionnaires take three to five business days because an engineer answers them rather than a template. For live client escalations, a P1 is acknowledged in 15 minutes under the standard support terms.

USD 12,000 for a fixed-scope discovery and prototype, four weeks for capped time and materials, or two engineers for three months on a dedicated team. An offshore development centre needs roughly ten engineers and a thirty-month horizon before the structure makes sense. Below those thresholds our overhead is a poor deal for you, and we will say so.

Yes, and about a third of our client work is early-stage, though we are candid about fit. We work well with funded startups that have a technical founder or a product owner able to make decisions weekly. We work badly with pre-revenue teams seeking equity-for-work arrangements, and we do not take them, because a bootstrapped firm cannot carry that risk.

We decline staffing-only contracts with no delivery visibility, lowest-bid procurement, fixed prices on genuinely unclear scope, timelines already impossible at signature, and engagements needing domain depth we do not have. We also do not build gambling platforms, and we do not take work requiring us to hold production credentials we should not have. Roughly one enquiry in five leaves with a smaller recommendation than it arrived with.

Fixed-scope work starts at USD 12,000 for a discovery and prototype. A typical enterprise web application lands between USD 60,000 and USD 220,000. A dedicated engineer costs from USD 6,500 per month depending on role and seniority. Most clients see a 45 to 60 percent reduction against the equivalent onshore build. Every proposal carries a fixed price or a capped ceiling before work starts.

Yes, where the scope genuinely supports it, and we carry the estimation risk: if the build takes 30 percent longer than we thought, that is our problem rather than your invoice. We will not fix a price on a system we have not inspected, so discovery is priced separately and first. Where scope is ambiguous we propose a capped time and materials ceiling instead.

We invoice in USD, GBP, EUR, AED or INR, by international wire over SWIFT or domestic transfer for INR. Standard terms are 15 days from invoice date, with 30 days available on request and commonly agreed with larger organisations. Team-based work is invoiced monthly in advance, fixed-scope work against milestones. We quote and invoice in one currency for the life of the contract.

You hear about it in the next stand-up, not in the retrospective. Slippage, a wrong architectural call or an underestimate is raised upward immediately, with a written recovery plan inside three working days covering what changed, the options and the cost of each. Escalation to a founder is a documented path with a response time attached. Where the error is ours, the remediation is not billed.

You give 30 days notice on team-based work, or stop at any sprint boundary with two weeks notice on capped time and materials, and pay only for work delivered. Exit assistance is a contractual deliverable: repository and infrastructure handover, runbooks, architecture decision records and a paid overlap period with your incoming team. There is no termination penalty and no data hostage situation, because your code lives in your repositories throughout.

Yes, under a separate support agreement with defined severity levels, or as part of a continuing dedicated-team arrangement. Standard cover includes a 15-minute P1 acknowledgement, a 99.95% availability target where we control the platform, and monthly reporting on incidents, patching and cost. Warranty defect remediation against agreed acceptance criteria is free for 30 days after each milestone regardless of whether you take support.

We commit 4–8 hrs of daily overlap in writing in the statement of work, varying by market, and staff shifts to hold it rather than treating it as best effort. UK clients get about seven hours, the Gulf about eight, the US East Coast about 4.5 in your morning. Call slots run 06:00 to 22:00 IST. No client is asked to take a 3 a.m. meeting.

Our teams take twelve public holidays a year, and the dates for the coming year are attached to every statement of work so they never surprise your plan. Support and on-call cover continues through all of them under support agreements. Where a holiday falls inside a cutover window we plan around it or staff it as paid cover, agreed in advance rather than discovered on the day.

Voluntary attrition runs under 9 percent annually against an Indian industry average of 18 to 25 percent, and average tenure on a client account is about 2.5 years. Contractually you get a named roster, a 90-day no-swap commitment and 30 days notice of any change after that. The structural reasons are our Dehradun base, published pay bands and engineers dealing with clients directly.

We replace them within two to three weeks and fund a two-week paid overlap so the incoming engineer learns your system from the outgoing one rather than from a wiki. Handover notes, architecture decision records and runbooks are maintained continuously for exactly this reason. If the departure is a performance replacement initiated by us, the overlap and the ramp are at our cost, not yours.

Yes, and the honest answer is that it costs money. Our contracts carry a non-solicitation clause for twelve months after an engineer leaves your account, with a buy-out fee of typically 40 to 50 percent of the engineer’s annual fixed pay if you want to convert them. We will not obstruct a hire someone genuinely wants. We do expect to be compensated for recruiting and training them.

No. Every engineer on your engagement is employed by us, works from our Dehradun or Bengaluru offices, and has signed an individual NDA and IP assignment before access is granted. We do not use contractor marketplaces or partner bodyshops. Where a genuinely specialist third party is needed, for example a penetration testing firm, we name them and get your written approval first.

You do, and it assigns to you on creation rather than on final payment, which is unusual and deliberate. Code lives in your repositories under your organisation from the first commit. Every engineer signs an individual IP assignment, not just a company-level clause. Where we reuse our own pre-existing libraries we tell you which ones and grant a perpetual, irrevocable licence rather than leaving a dependency you cannot use.

Yes to both, and we will sign yours rather than insisting on our template. NDAs are usually returned executed the same working day. Data processing agreements take two to five business days because we read them properly, and we will flag anything we cannot actually comply with rather than signing and hoping. We can also execute standard contractual clauses for transfers out of the UK or EEA.

Yes, and an engineer answers it rather than a bid template. Allow three to five business days for a full questionnaire. We also provide our information security policy set, subprocessor list, encryption and key-management detail, and breach-notification commitments. We are not currently certified to ISO/IEC 27001 or SOC 2, we say so plainly on every questionnaire, and our control set is mapped to ISO 27001 Annex A.

Wherever your contract specifies, and by default in your own cloud accounts in your own chosen region rather than ours. We work inside your tenancy with least-privilege, time-bounded access. Where we do host on your behalf, the default regions are London or Frankfurt for UK and EU clients, Bahrain or the UAE for Gulf clients, and Mumbai for Indian clients. Production data is not copied to India for development.

Least-privilege access with named individual accounts, hardware-backed multi-factor authentication on all administrative access, full-disk encryption on every endpoint, background verification before onboarding, mandatory code review, dependency and secret scanning in the pipeline, and centralised audit logging. Client production access is time-bounded and logged. The honest limit: we are a 40-plus person firm, so we do not run a 24-hour security operations centre in house.

Node.js with TypeScript, Python, Java and .NET on the backend. React and Next.js on the front end. React Native and Flutter for mobile. AWS and Azure for cloud, with Terraform, Kubernetes, ECS and Container Apps. PostgreSQL, MySQL, MongoDB, Redis. Snowflake, BigQuery, Redshift and Databricks for data, with dbt and Airflow. Amazon Bedrock, Azure OpenAI and open-weight models for AI work.

Yes, and that is the normal case. We work in your repositories, your cloud accounts, your CI system, your ticket tracker and your ceremonies. If your stack includes something we do not know, we say so and price the ramp honestly rather than proposing a rewrite that happens to match our comfort zone. We will tell you if we think a technology choice is a genuine problem, once, in writing.

Yes, as a reviewed assistant, and we disclose it in the statement of work. Engineers use licensed assistants on business plans configured so your code is never used for model training. Every generated line passes the same human review and test gates as hand-written code, and the reviewing engineer is accountable for it. Licence and provenance scanning runs in the pipeline, copyleft-risk suggestions are blocked, and client code never goes into a consumer chatbot. IP assignment to you is unaffected. If you want assistants prohibited, we will contract to that.

Automated tests are written with the feature rather than after it, and the pipeline blocks a merge on failing tests, coverage regression, linting, dependency vulnerabilities and exposed secrets. Playwright or Cypress covers critical user journeys, with API contract tests between services. A deployable increment lands in your environment fortnightly and is demoed live. Where a sprint produced nothing demonstrable, we say that plainly.

Yes, for the interfaces we build, covering user flows, wireframes, a component library and WCAG 2.2 AA accessibility. We do not do brand identity, marketing sites or campaign design, and we will tell you to hire a specialist studio for those. On enterprise applications, design effort is typically 10 to 15 percent of the engagement rather than a separate phase with its own sign-off gate.

More than 40 engineers across Dehradun and Bengaluru, serving clients in 10 countries, operating for 4 plus years since October 2022. Deliberately mid-sized: large enough to staff a fifteen-person programme, small enough that a founder is reachable on any escalation. We would rather grow slowly than dilute the delivery quality that produced 96% client retention.

Bootstrapped, with no external investment, and profitable each year since 2023. That has two consequences worth knowing. We can decline work that does not fit rather than chase a revenue target set by an investor timetable. And we cannot absorb a large unpaid receivable, which is why payment terms matter to us and why we ask for milestone payments rather than settling everything at the end.

Yes to both. Visitors are welcome at Dehradun or Bengaluru with a few days notice, and clients typically combine a visit with a two-day planning workshop. For travel to you, we hold valid business travel documentation for our senior team and will attend on-site for discovery, cutover or governance reviews. Travel and accommodation are charged at cost with no mark-up; our time on site is not charged extra.

A named delivery lead and a named lead engineer, both identified in the statement of work, plus the engineers on your roster. Nitesh, Head of Global Engagement, owns the commercial relationship. The architect who scoped and estimated your work is contractually the person who leads it, which is deliberately inconvenient for us because it stops anyone promising what the build team then has to absorb.

Yes, at proposal stage, matched to your sector and engagement shape, and we introduce you directly rather than sending a testimonial we wrote. Client retention is 96%. What we will not do is name clients publicly without written permission, which is why our case studies describe a UK insurance broker or a Gulf logistics operator rather than naming them. Reference calls are arranged within about a week.