Java · Node · Python · Go · Angular · React · AWS
New systems, and the ones that have outlived their stack. Architecture, build, migration — delivered as bounded modules, each with its own price, date and written acceptance criteria. No hourly billing, no scope creep, no overrun. The work decides the stack, not me.
Book a 20-minute call See pricing
The enterprise stack built around 2015 — a JVM or .NET backend, an AngularJS
front end, on-prem deployment — reached end of life on all three layers at
once. Most contractors can fix one. I quote the whole thing, so the
integration risk stays with me instead of with you.
Fifteen years of this across healthcare document systems, financial and
billing platforms, building operations, automation and integration work — so
I usually understand the domain before I open the codebase, which is where
this kind of work tends to go wrong.
Greenfield work where the scope is defined: an architecture, a sequenced plan, then the build in modules. Most useful where the system has to be right the first time rather than iterated into shape.
Out of support and blocking your security review. Mechanical work that touches every file — which is exactly the kind that overruns when it is billed hourly.
AngularJS has had no security patches since January 2022. Migrated incrementally, screen group by screen group — nothing goes dark mid-way.
WebLogic and Oracle licensing commonly runs $50k+ a year. The migration usually pays for itself inside twelve months.
Three named endpoints to an agreed p95, measured before and after on staging. Half the fee is contingent on hitting the target.
A week reading what you already have, then a written roadmap: what to change, in what order, and what each piece costs. No code, no commitment beyond the week, and the document is yours whether or not I build any of it.
Every engagement starts with a paid assessment. You get a fixed price on every module before you commit to any of them — and the assessment is useful on its own even if you stop there.
Three to five days. I read the codebase and return every blocker, the effort behind each one, a risk register, and a fixed price per module.
Take all of them, some of them, or none. Each module carries itemized deliverables, objective acceptance criteria and a delivery date.
Short gates, not quarterly ones. You review working code against the written criteria at every milestone, so nothing runs for months before you see it.
Tests green, documentation written, code transferred. For cloud work, your team runs the production cutover from my runbook, with me on the call.
Published, because a number you can plan against is the point. Everything is fixed-price — if the work takes longer than I estimated, that is my problem and not your budget's.
| Engagement | Scope | Price | Terms |
|---|---|---|---|
| Assessment | 3–5 days. Written report, blockers, effort, fixed module quotes | $900 – $2,500 | 100% upfront |
| Architecture roadmap | 1 week. Current state, target architecture, sequenced plan. No code | $2,500 – $5,000 | 100% upfront |
| Performance | 2–3 weeks. Three endpoints to an agreed p95 | $6,000 – $15,000 | 50% up, 50% on hitting the number |
| Single module | 2–4 weeks. One bounded slice of one layer | $5,000 – $12,000 | 50% / 50% |
| Multi-module | 5–8 weeks. Two layers | $15,000 – $28,000 | Milestone-gated |
| Full program | 10–14 weeks. Backend, frontend and cloud | $30,000 – $55,000 | Milestone-gated |
Payment due on invoice; work begins on receipt of funds. Intellectual property transfers on final payment. Two review rounds are included in every module — anything outside the written acceptance criteria is a change order with its own price and date.
Fifteen years of enterprise delivery. Client work is described without naming the client; the reference implementation is public and runnable.
Floor plans with devices plotted by position and coloured by status, per-floor alert handling, deviation from setpoint, and estate demand. A fault is located, not just listed.
A supplier catalogue for architectural lighting. Specification filtering on colour temperature, beam angle and ingress rating, with trade accounts, cart, checkout and order history.
A runnable reference implementation of the building-telemetry architecture I use: ingest, windowed anomaly detection, and command dispatch over Kafka.
Reads a folder of building-services handover documents and reports what is missing — the assets with no commissioning certificate, the certificates for plant that was never in the register, and the scans nobody can read.
A configurable engine for construction project delivery — pipelines defined as data, so routing changes without a deployment. Included extracting a 3,100-line route file into a service layer behind repository interfaces.
Event-driven monitoring and control across multiple office campuses — HVAC, electrical metering and DC power — decomposed so ingestion throughput and analytics load scale independently.
IRR modelling for commercial property deals, replacing the spreadsheets an investment team had been pricing deals with. Capex and opex drivers held as versioned data rather than cells.
Tell me what you are running and what is blocking you. I reply within one business day with scope and a price, or tell you straight away if it isn't a fit.
Or email directly: ashish@ashishaggarwal168.com
LinkedIn · GitHub