Software and AI engineering that holds up to scrutiny
Software development services from Nepal, with published team sizes, start times and notice terms. Add engineers to your team, hand us a scoped project, test an idea, or bring us in for AI and security work.
- 1 to 15+ engineers
- Engineers start in 5–10 working days
- 30 days' written notice
- Proposal within two working days
Engineers who work on your roadmap
Every model starts with people employed by us and run under the same ISO/IEC 20000-1 certified service management processes. The four models differ in who manages the work, how many engineers you need and for how long.
Dedicated teams
A team of 3–15 or more engineers working only on your product, usually a tech lead, developers and QA. You set the backlog and priorities; we handle hiring, payroll and the office. Best for long roadmaps. Ready in 2–4 weeks.
Read moreStaff augmentation
One to five engineers who join your existing team, attend your standups and work in your tools. You manage the work, we remain the employer. Suits a skills gap or extra capacity for a few months. Engineers start in 5–10 working days.
Read moreSoftware outsourcing
You give us a defined scope and we deliver it, with our own project management, milestones and acceptance criteria. Works when requirements are clear and you want a fixed outcome rather than a team to direct. Starts with a 1–3 week discovery.
Read moreAI staffing
Machine learning engineers, data scientists, data engineers and data annotators placed in your team on the staff augmentation or dedicated team model. For companies that already know what they want to build and need people with AI experience to build it.
Read moreTest an idea before you commit a budget
Both are small-squad builds that take 4–8 weeks. The difference is the question they answer.
PoC development
A proof of concept answers "can this work?" It tests the riskiest technical assumption, such as an integration, a data source or a model's accuracy, with as little surrounding product as possible. Useful before an investor meeting or an internal funding decision.
Read moreAI MVP development
An MVP answers "will people use this?" It's a small but real product with an AI feature at its core, built to go in front of users and collect feedback. We keep the scope tight and build it so the code can grow into the full product.
Read moreAI, machine learning and data engineering
From the data pipelines underneath to the LLM features on top. Every AI build is measured against an evaluation set, so you know how well it works before it goes live.
AI/ML engineering and model deployment
Custom machine learning models for prediction, classification, forecasting and vision, taken from notebook to a deployed, monitored service that your product calls.
Read moreData engineering and architecture
Pipelines, warehouses and data models that make your data reliable enough to report on and train models with. Usually the first job in any serious AI project.
Read moreMLOps and AI infrastructure
Deployment, versioning, monitoring and cost control for models in production, on AWS, Google Cloud or your own servers. For teams whose models work in testing but are hard to run.
Read moreGen AI and LLM solutions
Answers from your own documents (RAG), assistants inside your existing tools and document extraction, built with evaluation harnesses, guardrails and cost budgets.
Read moreAI agent development
LLM agents that look things up, call your APIs and complete multi-step tasks, with tight permissions, confirmation steps for risky actions and full audit logs.
Read moreData labelling
Text, image and audio annotation by trained annotators, with written guidelines and quality checks. Start with a sample batch so you can judge the quality first.
Read moreTrusted AI development partner
A long-term arrangement where we own the AI side of your product from scoping through operations, for companies without an in-house AI team.
Read moreSecurity, compliance and AI governance
Two related services. One covers your whole environment, the other the AI systems within it.
AI governance and security
AI system inventory and risk classification, EU AI Act and ISO/IEC 42001 readiness, NIST AI RMF programmes and LLM security testing, including prompt injection and red teaming. Starts with a 2–6 week scoping phase.
Read moreCybersecurity and GRC
Penetration testing, ISO 27001 and SOC 2 readiness, phishing simulation, security training and vCISO support. For companies preparing for an audit, a customer security review or a regulator.
Read moreHow do Dignep's engagement models compare, and how is pricing set?
We don't publish rates because every quote depends on the work. Pricing is scoped after a free 30-minute discovery call, and you get a written proposal within two working days. What drives the number: seniority mix, team size, engagement length, technology stack and, for governance work, compliance scope.
| Model | Team size | Start time | Notice or end | Best fit |
|---|---|---|---|---|
| Staff augmentation | 1–5 engineers | 5–10 working days | 30 days' written notice | Extra capacity in a team you manage |
| Dedicated team | 3–15+ engineers | 2–4 weeks | 30 days' written notice | Long roadmap, full-time focus |
| Project-based | Scoped | 1–3 week discovery | Milestone acceptance | Defined scope, fixed outcome |
| PoC or MVP | Small squad | 4–8 week build | Milestone acceptance | Testing an idea before a bigger investment |
| AI governance or GRC | Scoped | 2–6 week scoping | Milestone acceptance | Regulation, audits, AI risk |
How pricing works: every engagement is scoped and quoted after a free 30-minute discovery call, and you get a written proposal with pricing within two working days.
Which model fits your situation?
Choose by how settled your scope is and how long the work will last. Picking the wrong one is the most expensive mistake we see, so here's how we'd advise a friend.
| Your situation | What we'd suggest | Why |
|---|---|---|
| You have an engineering manager and need two more React developers for six months | Staff augmentation | Your process stays the same. A dedicated team would add a layer you don't need. |
| You have a product with a year or more of roadmap and not enough engineers | Dedicated team | The team builds up knowledge of your product, and you avoid paying for onboarding again every few months. |
| You need a specific system built to a clear spec by a fixed date | Project-based outsourcing | You buy an outcome, not hours, and we carry the delivery risk on scope we've agreed. |
| You're not sure an idea is technically possible, or whether users want it | PoC or AI MVP | A 4–8 week build that ends in "no" costs far less than building the full product first. |
| A customer, auditor or regulator is asking how you control your AI | AI governance and security | The scoping phase gives you an inventory and gap list you can show them. |
When we're not the right fit: if you need one developer for two weeks, a freelancer platform will be quicker. If your team works US hours and needs constant real-time collaboration, our lack of natural overlap with US time zones will hurt, and we'd rather tell you now.
How does an engagement start?
With a free 30-minute call. We ask what you're building, what's slowing you down and what you've tried. If we're a good fit, you get a scoped proposal within two working days. If we're not, we'll say so on the call.
What we work in
Product engineering: React, Next.js, TypeScript, Node.js, Python, Supabase, Google Cloud, AWS.
AI: LLM agents, retrieval-augmented generation (RAG), evaluation harnesses, prompt engineering for production.
Governance and compliance: ISO 27001, SOC 2, EU AI Act, NIST AI RMF, Nepal Rastra Bank AI guidance.
Discovery call
Thirty minutes to talk through what you need. Expect us to ask more questions than we answer.
Scoped proposal
Within two working days: the model we recommend, team or scope, price, start date and the risks we see.
Working hours agreed
We set the daily overlap window and communication tools before kickoff, using the time-zone table below.
Kickoff
On the start date in your proposal, with the start times shown in the table above.
Working across time zones
Our day runs 09:00–18:00 Nepal Time (UTC+5:45, no daylight saving). India shares almost the whole day, Central Europe about four hours, Sydney about three to four and the UK about three. The US has no natural overlap, so we agree a fixed daily call window before kickoff, usually by moving the team lead's working day.
Work you can check
We've delivered more than 100 projects, including public-sector work where audits and reporting are part of the job. A few to start with: the Certifyi GRC platform, the SayCure SOC platform, the Real-Time Monitoring System for the Town Development Fund and our work on the UNDP Innovation Partnership Fund. Our applied research happens in DAIR, working with Kathmandu University's Department of Artificial Intelligence. See all case studies.
Questions about working with Dignep
What is the difference between a dedicated team and staff augmentation?
A dedicated team is a group of 3–15 or more engineers working only on your product, usually with its own tech lead. Staff augmentation adds 1–5 individual engineers to a team you already manage. Both stay on our payroll and both need 30 days' written notice to end.
How quickly can you start?
Staff augmentation engineers usually start in 5–10 working days. A dedicated team takes 2–4 weeks to put together. Project work begins with a 1–3 week discovery, and a PoC or MVP build runs 4–8 weeks. Every engagement starts with a free 30-minute call and a scoped proposal within two working days.
How does an engagement end?
Dedicated team and staff augmentation contracts end with 30 days' written notice. Project, PoC and MVP work ends on milestone acceptance, when you sign off the agreed deliverables.
Can we start small before committing to a team?
Yes. A PoC or MVP (a 4–8 week build) or a scoped project with a 1–3 week discovery are both reasonable ways to see how we work before you commit to a dedicated team.
Not sure which service you need?
Book a 30-minute call. Tell us what you're trying to ship and by when, and we'll recommend a model, including a smaller one if that's what makes sense.
