I build software your business can actually run on.
I’m Deepak Joshi — ex-Deloitte USI, founder of Dharmarthlabs, and the engineer behind Code Hippies. Full-stack web, iOS and Android apps, and AI systems that answer from your data instead of inventing answers. Thirteen sites shipped to production, and every technical claim about them was read off the live response — so you can check the work rather than take my word for it.
- sites shipped to production
- 13
- 12 are reachable right now and independently verifiable — links below.
- running on Next.js
- 4
- Confirmed by x-powered-by response headers on the live origins.
- shipping structured data
- 10
- Schema.org JSON-LD emitted in the served HTML, not injected client-side.
- lightest page shipped
- 21 KB
- nantinbaba.org — static HTML, no framework runtime, no maintenance contract.
Not sure what you need yet? Read the FAQ or see how engagements are priced.
Live production sites built by Code Hippies
- Uttaranchal Kesari — uttaranchalkesari.com
- Himachal Kesari — himachalkesari.com
- Haryana Kesari — haryanakesari.com
- NewsLive24 — newslive24.in
- ReadyNews — readynews.in
- IndiaNews16 — indianews16.com
- CarbonMedia — carbonmedia.in
- FitWithNash — Consultation — consult.fitwithnash.com
- Core Media Solutions — coremediasolutions.in
- Influence Axis — influenceaxis.in
- ScaleWell Digital Solutions — scalewelldigitalsolutions.in
- Belong Digital Solutions — belongdigitalsolutions.in
Start where you are
Same work. Three ways of explaining it.
You might be running a shop and dreading being talked down to. You might read diffs for a living. You might be signing off a budget that was going to fund headcount. Pick your version — the claims underneath are identical, and every one of them is checkable.
You know what your business needs. You do not know, and should not have to know, how it gets built.
You describe the problem. I handle everything with a screen on it.
Most people who need software have been burned once already — a developer who went quiet, a site that broke and nobody could fix, an invoice that doubled. So here is how this works in plain terms, with no jargon and nothing hidden.
You get a price before I start, and it does not move
We talk first, I write down exactly what you are getting, and I quote a fixed number against that. If you later want something extra, I tell you what it costs before I build it. You will never open an invoice and find a surprise.
You can see it being built, on your own phone
From the first week there is a live link you can open any time. You are not waiting until the end to find out whether I understood you. If something looks wrong, you say so while it is still cheap to change.
It is yours, and it keeps working after I leave
The site or app is registered in your name, on your accounts, with your card. I hand over instructions written so that any other developer can pick it up. You are never stuck with me because you cannot leave.
“How do I know you will not disappear like the last one?”
Because everything is in your name from day one — the domain, the accounts, the code. Thirteen sites I have built are live right now and you can click every one of them. And if you want the safety of a platform that holds your money in escrow until you are happy, I work on Fiverr and Upwork too. You do not have to trust me on faith.
You can read a diff and a P&L. You are assessing judgement, not vocabulary.
The interesting question is not what I can build. It is what I will talk you out of.
You already know a competent engineer can ship most things. What separates a good engagement from an expensive one is whether the person you hire will tell you when the interesting solution is the wrong one. So here is the evidence, not the pitch.
Six stacks in one portfolio, on purpose
Next.js for regional news properties that need stories crawlable in seconds. WordPress left in place for a newsroom that would have lost weeks migrating. 21 KB of static HTML for an ashram that cannot carry a maintenance contract. Each of those was the cheaper answer for that constraint, and two of them were less interesting to build.
The site is the work sample, and it is falsifiable
Every technology claim on the case studies was read off the live response headers — you can re-run the curl yourself, the method is in the repo. This site holds 100/100/100/100 on Lighthouse desktop with a budget enforced in CI, zero WCAG 2.1 AA violations, and an AI assistant that refuses out-of-scope questions with the refusal path tested on every deploy.
AI where it earns its place, not where it demos well
Retrieval-grounded systems with a relevance floor, an explicit refusal path, and a golden question set in CI where half the cases must be declined. I will not sell you a guarantee of zero hallucination, because nobody can deliver one. What I will do is make the failure mode a visible refusal instead of a confident invention, and prove it with a test you can read.
“Why you instead of an agency with twenty people?”
For a defined product build, twenty people is mostly coordination overhead you are paying for. One senior engineer who scopes the work, writes it, and owns the outcome removes an entire communication layer. Where scale genuinely helps — parallel workstreams, a hard deadline, ongoing multi-product delivery — that is what the Dharmarthlabs partnership exists for, and I will point you there rather than pretend I am a team.
You have a budget line for headcount and a delivery date you do not control.
A senior engineer who ships, without the eighteen-month hiring cycle.
Replacing a team is the wrong frame, and I will not sell it that way. What actually happens is narrower and more useful: for a defined scope, one accountable senior engineer with modern tooling delivers what a small team used to, at a fraction of the coordination cost — and for anything beyond that scope, there is a partner network to escalate into rather than a solo bottleneck.
What one engineer plus modern tooling genuinely covers
A production web application, a cross-platform mobile release, a retrieval-grounded AI feature with an evaluation harness, a CI/CD pipeline with performance and security gates, and the technical SEO underneath all of it. That was a four-to-six person scope five years ago. It is not any more, and pretending otherwise costs you money.
What it does not cover, said plainly
24/7 on-call rotation. Multiple parallel product lines. A team that survives one person going on holiday. If you need those, you need a team or a partner — and that is exactly the conversation Dharmarthlabs exists for. An engineer who tells you this before the contract is worth more than one who finds out during it.
Governance you can put in front of a procurement review
Ex-Deloitte USI, which is where I learned what production actually costs: review gates, audit trails, and a change nobody can trace being a change nobody can defend. Every engagement ships with a documented pipeline, an OWASP Top 10 review, dependency scanning that fails the build, and handover written so your next hire needs fifteen minutes, not a meeting with me.
“What is our exposure if you are unavailable?”
Contained by design, and worth checking rather than trusting. The repository sits in your organisation with full history, every credential is in your accounts, the pipeline runs without me, and the handover gets a new developer running locally in under fifteen minutes. Where continuity is a formal requirement, the Dharmarthlabs partnership provides a named backup team under the same agreement.
How I work now
I learned the slow way first. That is what makes the fast way safe.
Anyone can generate a website this year. Far fewer people can tell you whether the thing that came out is correct, secure, accessible, and going to survive its second year in production. The fundamentals are not nostalgia — they are the review layer that makes modern tooling worth using.
The foundation
Learned first- Semantic HTML and the document outline — still what search engines and screen readers actually read
- CSS layout from first principles, so a broken grid is diagnosed rather than guessed at
- Relational data modelling and query plans, because the slow page is usually the query
- HTTP, caching and the request lifecycle — the layer most performance problems actually live in
- OWASP Top 10 and defensive server-side validation as a habit, not a checklist
The modern stack
Used daily- React Server Components and edge rendering, chosen per route rather than by default
- TypeScript in strict mode with validation schemas shared across the client/server boundary
- Retrieval-grounded LLM systems with evaluation harnesses that run in CI
- AI-assisted development held to the same review bar as anything else — generated code is reviewed, not trusted
- Infrastructure as configuration, with performance and security budgets that fail the build
The combination is the point. Fundamentals without modern tooling is slow. Modern tooling without fundamentals is fast, confident and wrong — and you will not find out until it is in production.
What I do
Five things, done properly, instead of everything done adequately
Each of these is a distinct offer with its own scope, its own deliverables and its own way of being judged. Pick the one that matches your problem.
Web development
Server-rendered web applications and marketing sites built on Next.js and React — typed end to end, indexable by default, and deployed through a pipeline that will not let a broken build reach users.
Explore this serviceiOS & Android development
iOS and Android applications — native Swift and Kotlin where the platform matters, React Native or Flutter where shared code matters more — taken through store review and into a release pipeline you can run yourself.
Explore this serviceAI & LLM engineering
RAG systems, AI assistants and LLM-backed workflows built so the model answers from retrieved source documents and says it does not know when nothing relevant is retrieved — with the evaluation harness to prove it.
Explore this serviceSEO & performance engineering
Structured data, rendering strategy, crawl architecture and a Core Web Vitals budget enforced in CI — the technical half of SEO, which is the half a developer is actually responsible for.
Explore this serviceSecurity & compliance consulting
Application security review against the OWASP Top 10, security headers, dependency scanning in CI, and straight guidance on what compliance work you can automate and what genuinely requires auditors and money.
Explore this service
Selected work
Thirteen live sites. Open any of them right now.
Each case study pairs what was actually engineered with what it does for the business — and lists the exact signals observed on the live response, so nothing here needs to be taken on trust.
- News & Publishing2026
Uttaranchal Kesari
EngineeringNext.js application server (confirmed by an
x-powered-by: Next.jsresponse header) fronted by Caddy for TLS termination, serving server-rendered Hindi article markup atlang="hi"withNewsMediaOrganization,WebSiteandSearchActionJSON-LD emitted on the document.In plain termsPublishes Uttarakhand's daily Hindi news in a form Google News and search engines can read the moment a story goes live.
- Next.js
- React
- Caddy
- Server-side rendering
Read the case studyuttaranchalkesari.com(opens the live site in a new tab) - Health & Coaching2026
FitWithNash — Consultation
EngineeringAstro site (public source at github.com/codehippies11/fitwithnash) deployed to Vercel's edge network — the served page is essentially all content and almost no asset weight, and it carries the deepest structured-data graph in the portfolio: Person, ProfessionalService, Service, OfferCatalog, Offer, EducationalOccupationalCredential, Audience and a full FAQPage.
In plain termsTurns a nutritionist's practice into a single page that answers every question a prospective client has, then books them.
- Astro
- Vercel
- Static generation
- JSON-LD / Schema.org
Read the case studyconsult.fitwithnash.com(opens the live site in a new tab) - News & Publishing2025
NewsLive24
EngineeringWordPress origin behind nginx, instrumented with GA4 via Google Tag Manager and monetised with Google AdSense; the document carries the richest structured-data graph of the portfolio — NewsArticle, NewsMediaOrganization, CollectionPage, ItemList, ContactPoint and SearchAction — plus hreflang alternates.
In plain termsRuns a full English news desk on a CMS the editors already know, while the technical SEO and ad plumbing is handled underneath them.
- WordPress
- nginx
- GA4 / Tag Manager
- Google AdSense
Read the case studynewslive24.in(opens the live site in a new tab)
How it runs
You always know what happens next
Nine stages from discovery to maintenance. Every one has a defined output and a defined thing I need from you — because the projects that go wrong are almost always the ones where nobody wrote that down.
- 01
Discovery
3–5 daysBefore anything is estimated, we establish what the software has to do, who it is for, and what makes it a failure. Most bad projects were mis-scoped, not badly coded.
- 02
Planning & wireframing
3–7 daysScreens and data model are decided on paper, where changing your mind is free. Every screen is mapped to the data it needs and the actions it allows.
- 03
Design
1–2 weeksA design system rather than a set of pictures — tokens, type scale, spacing, states. Components are designed with their loading, empty and error states, because those are most of what users actually see.
- 04
MVP
2–3 weeksThe thinnest version that a real user can complete a real task on, deployed somewhere you can open on your phone. Not a demo — a deployment.
- 05
Development
3–10 weeksFull build in increments, each one merged behind a passing pipeline. You see progress on a preview URL continuously rather than at a milestone meeting.
Evidence
Verifiable, not testimonial
The strongest proof an engineer can offer is work you can inspect without asking permission.
There are no client quotes on this page yet, and I would rather show you something you can check than something you have to take my word for. Every claim below was read directly off the live site’s response — open any of them and verify it yourself.
- uttaranchalkesari.com(opens in a new tab)
- Response header
x-powered-by: Next.js - Response header
server: Caddy <html lang="hi">with server-rendered Devanagari headlines in the initial HTML
- Response header
- himachalkesari.com(opens in a new tab)
- Response header
x-powered-by: Next.js - Response header
server: cloudflare— CDN in front of the origin <html lang="hi">; server-rendered section headings such as मुख्य समाचार present in source
- Response header
- haryanakesari.com(opens in a new tab)
- Response header
x-powered-by: Next.js - Response header
server: Caddy <html lang="hi">; homepage renders an explicit empty state rather than an error
- Response header
- newslive24.in(opens in a new tab)
- Response header
Server: nginx - 71 references to
wp-contentin the served HTML — WordPress origin - Google Tag Manager / gtag GA4 loader present
- Response header
- readynews.in(opens in a new tab)
- Single application bundle
/assets/js/app.js?v=…with version-pinned cache busting - No CMS signature in the served HTML (no wp-content, no generator meta)
- Cloudflare references in the document
- Single application bundle
- indianews16.com(opens in a new tab)
- Application bundle
/assets/js/app.js?v=20260705atlas1— dated release tagging - 90 anchor elements and 28 images in the server-rendered homepage HTML
- Cloudflare references in the document
- Application bundle
Engagement
Three ways to work together
Fixed-scope projects, monthly retainers and staff augmentation. Pricing follows discovery, because a number quoted before the scope is understood is wrong in one direction or the other.
Bigger than one engineer?
There is a team behind this when you need one
Some work genuinely outgrows a single contractor — parallel workstreams, a productised build that should ship in days rather than months, or a technical partner with real stake in the outcome. That is Dharmarthlabs, the company I founded. Same person, with a team behind him.
You want a standard product, configured — not a bespoke build
Buy it productised
Most businesses asking for custom software need something that has been built four hundred times before: a booking flow, a news portal, a lead-capture site, a storefront, an internal dashboard. Dharmarthlabs builds those once, properly, and configures them per client. You pick the options, it ships in days rather than months, and you pay a fraction of a bespoke build because you are not funding someone's first attempt.
- Ships in days, not months — the product already exists
- Priced as a product, not as an open-ended engagement
- Options selected up front, so the scope conversation is a form rather than a workshop
- Same engineering standards as a custom build, because it is the same people
You are a startup that needs engineering leadership, not a vendor
Take on a technical partner
Startups without a technical founder usually get one of two bad outcomes: an agency that bills for a product it has no stake in, or a junior hire carrying architectural decisions they have never made before. The partnership route puts a technical team alongside the business with real skin in the game — including equity-inclusive structures where the fit and the stage justify it.
- Technical direction and architecture ownership, not just delivery
- Cash, equity, or blended structures depending on stage and fit
- A team that scales with you rather than a contract you renegotiate
- Diligence-ready documentation from the start, because investors will ask
FAQ
The questions that come up in every first call
Six of them here. The other 44 are on the FAQ page and inside the assistant in the corner, which you can browse by category or just ask.
Who is Deepak Joshi and what is Code Hippies?
Deepak Joshi is a full-stack, mobile and AI/LLM engineer who previously worked at Deloitte USI and is the founder of Dharmarthlabs. Code Hippies is the studio he builds under — an independent engineering practice serving startups and agencies, not a reseller and not a body shop. Every project is delivered by the person you talk to in the first call.
Can you take over a project another developer started?
Yes, and it is common. It starts with a short audit — what state the code is in, what is salvageable, what is a rewrite — delivered as a written assessment before any code changes. Sometimes the honest answer is that continuing costs more than restarting, and I will tell you that.
How much does a website or app cost?
A focused marketing or consultation site is typically two to four weeks of work. A full web application with authentication, a database and an admin surface is usually six to twelve weeks. A first mobile release is eight to fourteen weeks. Exact pricing comes after discovery, because quoting before understanding the scope produces a number that is wrong in one direction or the other.
Do you work fixed-price or hourly?
Fixed-scope projects are fixed-price, quoted after discovery against a scope you have signed off. Ongoing work runs as a monthly retainer with a reserved number of days. Joining an existing team runs on a day rate. Hourly billing on an undefined scope serves neither of us.
What technologies do you build with?
Web on React/Next.js, Vue/Nuxt or Astro; backends on Node/NestJS, Python/Django or FastAPI, Go, or PHP 8.2+; mobile on Swift and SwiftUI for iOS, Kotlin for Android, or React Native and Flutter for cross-platform. The stack is chosen per project against your constraints — not every project uses every one of these.
Who owns the code you write?
You do, from the first commit. Work happens in your repository or is transferred to your organisation at handover, with full history, environment documentation and a local setup a new developer can run in under fifteen minutes.
Question not covered here? The assistant in the corner has all 50 of these grouped by category and searchable, answers anything else from this site’s knowledge base, and tells you honestly when it does not know. Or send the brief directly.
Tell me what you're building.
Two minutes on the brief form gets you a real conversation: what the software has to do, what it will take, and an honest read on feasibility — whether or not we end up working together.