
We build the software your business actually runs on.
A three-person freelance engineering studio shipping production web applications, AI systems and automation — with fixed scope, honest timelines, and direct access to the people writing your code.
- Next.js
- Node
- Python
- PostgreSQL
- AWS
- Replies within 4 hours
- Fixed-scope quotes
- NDA on request
- 0+
- Projects shipped
- 0+
- Clients served
- 0+
- Years building
- 0%
- On-time delivery
Things we built that are still running.
Every project below shipped to production and stayed there. Pick one to see the stack, the timeline and what it changed for the client.
Placeholder content. These six projects, their clients and their results are invented examples. Replace it from the admin dashboard before launch — these notices vanish automatically once real records exist.
Everything between an idea and a running product.
We take work end to end — design, build, deploy and maintain. If a project needs something outside this list, ask; we will tell you honestly whether we are the right people for it.
Website Development
Fast, secure, accessible sites built on Next.js and React.
- Responsive
- SEO-ready
- Core Web Vitals
Landing Pages
Conversion-focused pages that turn ad spend into qualified leads.
- Clear CTAs
- A/B ready
- Analytics wired
Business Websites
Corporate sites that establish credibility and are easy to update.
- CMS included
- Secure hosting
- Lead capture
Portfolio Websites
Showcase sites for studios and individuals who are judged on craft.
- Custom motion
- Case studies
- CMS included
E-Commerce
Storefronts with custom checkout, subscriptions and an ops dashboard.
- Stripe
- Order admin
- Inventory sync
AI Chatbots
Assistants grounded in your own docs, with clean handover to a human.
- RAG retrieval
- Context memory
- Human handover
Machine Learning
Forecasting, scoring and segmentation models trained on your data.
- Python
- Pipelines
- Served via API
Deep Learning
Vision and language models for classification and extraction work.
- CNNs & RNNs
- Transformers
- OCR pipelines
Python Automation
Scrapers, report generators and back-office scripts that run themselves.
- Web scraping
- Scheduled jobs
- API syncs
API Development
REST and GraphQL services with auth, rate limiting and real documentation.
- JWT & OAuth
- OpenAPI docs
- Rate limiting
Dashboards
Internal consoles that make your operational data legible at a glance.
- Live data
- CSV export
- Role-based access
UI/UX Design
Interface design in Figma, delivered as a build-ready component system.
- Prototypes
- Design tokens
- Handover specs
Graphic Design
Brand marks, decks and social assets that stay consistent everywhere.
- Vector assets
- Brand guide
- Social kits
SEO Optimization
Technical audits, schema and speed work aimed at rankings that hold.
- Site audit
- Schema markup
- Rank tracking
Hosting & Deployment
Vercel, Render, AWS or Docker with CI/CD and automated backups.
- GitHub Actions
- Zero downtime
- SSL & backups
Maintenance
Ongoing patching, monitoring and small changes on a monthly retainer.
- Daily backups
- Dependency patches
- Uptime alerts
Not sure which of these you need? Describe the problem and we will scope it for you.
A process designed so you are never guessing.
Six stages, each with something concrete at the end of it. You always know what is happening, what it costs and when it lands.
- 01
Discovery call
30 minutes, freeWe talk through what you are trying to achieve, what already exists, and what success looks like. If we are not the right fit we say so on this call.
You receive
A written scope and a fixed quote — no obligation - 02
Architecture & plan
3–5 daysWe map the data model, pick the stack, identify the risky parts early and break the work into milestones you can track.
You receive
Technical plan with dated milestones - 03
Interface design
1–2 weeksWireframes first, then high-fidelity screens in Figma. You review and sign off before a single component gets built.
You receive
Clickable Figma prototype - 04
Build
2–10 weeksWe work in weekly increments against the milestone plan. You get a staging URL from week one, so progress is visible rather than described.
You receive
Weekly demo build on a live staging link - 05
Test & harden
3–7 daysAuth flows, file uploads and API limits get exercised properly. We run performance and accessibility passes before anything reaches your users.
You receive
Test report and security checklist - 06
Launch & support
OngoingWe deploy, monitor the first weeks closely and hand over documentation. Retainers are available if you want us on call after that.
You receive
Deployed app, docs and a handover session
Three engineers, not an agency.
Pinaki is a small freelance studio. We started it because good engineering was either locked behind agency retainers or delivered slowly by shops that treated the front end as an afterthought. We do both halves properly, and there is no layer between you and us.



You talk to the builders
No account managers relaying messages. The person answering your question is the person writing the code.
Scope is fixed before we start
You get a written scope and a fixed quote up front. If the work grows, we re-quote rather than quietly billing more.
Small book, on purpose
We take a limited number of projects at a time so nothing sits in a queue waiting for attention.
You own everything
Source code, design files and infrastructure are yours on final payment. No lock-in, no hostage situations.
What clients said afterwards.
Placeholder content. These reviews, the people quoted and their photographs are invented. Replace it from the admin dashboard before launch — these notices vanish automatically once real records exist.
“The CRM they built cut our inbox triage down to almost nothing. Churn predictions land around 92% accuracy and support satisfaction is up 30%. They shipped when they said they would.”
“Our landing pages convert at 18.5% now, up from under 7%. What impressed me most was that they pushed back on two ideas of mine that would not have worked.”
“The scrapers and lead bots replaced what used to be a four-person job. Deployed on AWS and running without intervention for eight months.”
Three ways to work with us.
We do not publish fixed price lists, because a number without a scope is meaningless. What we do publish is exactly how billing works, so there are no surprises on the invoice.
Fixed-scope project
A defined build with a clear finish line — a site, an app, an integration.
Quoted as one number after the discovery call
- Written scope agreed before work starts
- Milestone payments, not one lump sum
- Weekly demo builds on a staging URL
- Fixed price — scope changes get re-quoted
- 30 days of post-launch bug fixes
Monthly retainer
Ongoing feature work, maintenance and support on a product that keeps moving.
Reserved capacity billed monthly, cancel with 30 days notice
- A set number of days reserved each month
- Priority response ahead of project work
- Dependency patching and uptime monitoring
- Monthly written summary of work done
- Unused days roll over once
Consulting & audits
A second opinion — performance audits, architecture reviews, hiring support.
Hourly or day rate, invoiced on completion
- Codebase and architecture review
- Performance and Core Web Vitals audit
- Security and dependency assessment
- Written findings with prioritised fixes
- Follow-up call to walk through it
Not sure which fits? Send over what you have — a brief, a sketch, or just a paragraph describing the problem. We reply within 4 hours with a recommendation and a number, whether or not you go ahead with us.
Questions we get asked.
Anything not covered here, just ask. We reply within 4 hours and we would rather answer ten questions up front than have you commit to the wrong thing.
A marketing site or landing page is usually 2–4 weeks. A full application with a database, auth and an admin panel runs 6–12 weeks. You get a dated milestone plan before we start, so the estimate is not a guess you have to trust.
Notes from the build.
Occasional write-ups on decisions we had to make on real projects.
Placeholder content. These two posts are placeholder drafts. Replace it from the admin dashboard before launch — these notices vanish automatically once real records exist.
Scroll-linked animation in the Next.js App Router
Where to put motion so it survives streaming and server components, and which properties to animate if you care about frame budget.
Serving Python ML models behind a Node API
Two patterns for putting a trained model in production — a separate inference service versus a spawned subprocess — and when each one is the right call.
Tell us what you're building.
A paragraph is enough to start. We will come back with questions, a recommended approach and a fixed quote — free, and with nothing attached to it.
We deliberately run a small book so every project gets senior attention. If the month is full we will tell you straight away rather than stringing you along.