Skip to content
Devendra Jangiddevendra.pro

I speak both business and technology — and the gap between them is where projects fail

Business people and technical people are usually both right, and almost never talking about the same thing. I've spent most of my career in between the two — turning how a business really works into working software, and turning technical trade-offs into decisions an owner can actually make.

At a glance

Business owners trained & supported
100+
Years across ERP, product & operations
12+
ERP & automation rollouts delivered
40+
Industries served end to end
9
ERPNextOdooPostgreSQLPythonRSQLNext.jsFastAPIn8nRaspberry Pi

How I ended up doing this

I didn't plan on becoming an ERP consultant. I started on the technical side, building and shipping software end to end — which meant learning servers, databases, integrations, and what happens at 2am when something fails quietly and nobody notices until morning.

What changed my direction was watching businesses spend serious money on software nobody ended up using. The software itself usually worked fine. The problem was always earlier: nobody had written down how the business actually ran, nobody was responsible for getting staff to use the new thing, and the whole project had been handed to IT as if it were a purchase rather than a change in how people work.

So I learned the other half properly — operations, buying, how inventory really costs you money, what month-end close involves, and why a warehouse manager keeps his own private register. I kept the technical depth too. That combination is rare, and it's the whole basis of how I work.

Teaching, not just building

Along the way I kept having the same conversations. How do I judge this proposal? Is this quote fair? How do I hold a developer to account when I can't read code? Should we build this or just buy something?

Those questions don't need a consultant. They need someone to teach you how to answer them yourself. So that's what I started doing — sessions with owners, hands-on training for their teams, and staying available for advice afterwards. Over a hundred owners later, this is as big a part of my work as building systems, and honestly the part that lasts longest. A good system helps for five years. An owner who can judge technology decisions is set for the rest of their career.

Bringing businesses together

The third part grew out of the second. Once you've worked with enough small businesses, the waste becomes obvious: sixty independent shops each negotiating alone, each paying full price for software, each rebuilding the same thing from scratch.

Bringing them together fixes that — buying together, sharing one platform, sharing distribution. But it only works if the maths is proven before anyone builds anything, and if the rules are agreed in writing before the first argument. I design, build and run those groups.

What I'm not

I don't sell software licences, so I earn nothing from recommending one product over another. I don't supply developers by the month. And I'm the wrong person if you want someone to agree with a decision you've already made — I'll tell you what I actually think, which is occasionally inconvenient.

Where I can help

Right now

What I'm building this year

Client names stay private, so here is the work itself. Three live projects — a head-office platform, an AI automation system, and a legacy ERP being rebuilt from scratch.

Pharmacy retail · multi-state

Head-office operations platform for a pharmacy retail chain

The system the head office runs on. New store openings, the statutory licences each site needs, vendor quotations and purchase orders, recurring utility bills, the asset register and an internal ticket desk — all in one place instead of six inboxes.

  • New-site onboarding tracked against a full statutory checklist
  • Vendor quotation and purchase order flow with approvals
  • Recurring bills generated automatically each cycle
  • Role-based access with two-factor sign-in
Next.jsPostgreSQLServerless deployment
Pharmacy retail · AI automation

AI and automation platform for a retail pharmacy group

A working example of agentic AI doing real back-office work. A chat assistant store teams actually use, daily cash reconciliation, purchase margin and discount checking, and scheduled reports that build themselves overnight.

  • Chat assistant handling routine store and back-office questions
  • Daily cash reconciliation across the store network
  • Purchase margin and discount engine that verifies what was earned
  • AI agents taking over repetitive daily tasks
PythonFastAPIPostgreSQLAI agents
Legacy modernisation

Reverse engineering a legacy ERP into a modern web system

An old desktop ERP holding decades of business logic, with no documentation and no vendor support worth having. I mapped the entire system into a written blueprint, then began rebuilding it as modern web software — keeping every rule the business depends on, and none of the original code.

  • A very large legacy database mapped module by module
  • Full requirements blueprint written before any rebuild started
  • Clean-room rebuild: requirements copied, never code or schema
  • Modern web stack with layered role and branch permissions
NestJSPostgreSQLReactClean-room rebuild
The toolkit

What I actually work with

I pick tools per job rather than by habit. This is what I reach for most, and what I have running in production today.

Data & analytics

  • R (dplyr, tidyr)
  • Python
  • SQL & PostgreSQL
  • Databricks
  • Scheduled ETL jobs

ERP & business systems

  • ERPNext
  • Odoo
  • Frappe
  • Custom ERP builds
  • Legacy system rebuilds

Building & shipping

  • Next.js & React
  • FastAPI & NestJS
  • PostgreSQL
  • Docker
  • Cloud & serverless

AI & hardware

  • Whisper (STT)
  • Piper (TTS)
  • Local LLMs
  • n8n & AI agents
  • Raspberry Pi

A fair bit of it is public — have a look at my GitHub.

How I work

Four rules I stick to

Understand the work before touching software

Building before understanding is the most common reason projects fail. Two weeks of watching how work really happens changes everything that comes after.

Nobody pays me to recommend them

I earn nothing from recommending any particular software. I gave that money up on purpose — it is the only thing that makes my advice worth listening to.

You should end up not needing me

Making you dependent on me would be good business and a bad deal for you. Every project ends with guides, trained people and a team that can carry on alone.

Real numbers, not big stories

Time the manual work. Price the mistakes. Look at five years, not one. Most arguments about technology disappear the moment someone puts real numbers on the table.

Start with a conversation, not a proposal

Thirty minutes. Tell me what's going wrong and I'll tell you what I think — including if the answer is that you don't need to spend anything yet.

CallWhatsAppEnquire