Professional work

HashMicro ERP

A cloud-based business platform that connects finance, sales, inventory, procurement, and HR into one operational system.

Role
Python Programmer - ERP Business Workflows
Status
Archived
Year
2024
Stack
Python

What this project shows

Context

Product engineering on a cloud-based ERP platform connecting core business functions.

My contribution

Python development on the platform's business workflows across finance, sales, inventory, procurement, and HR operations.

Why it matters

Early production experience in ERP product engineering and long-lived business software.

Public evidence

Highlights

  • Python development on a cloud-based ERP platform
  • Business workflows spanning finance, sales, inventory, procurement, and HR
  • Modular business applications over a shared data model
  • Regional accounting, tax, and reporting requirements
  • ERP product engineering rather than one-off application code

ERP Product Engineering - Python - Business Workflows - Cloud Business Platform

Overview

HashMicro ERP is a cloud-based business platform that connects finance, sales, inventory, procurement, HR, and adjacent operational workflows into a single system. Publicly, it is described as a modular enterprise resource planning suite deployed in the cloud with an on-premise option, and localized for the accounting, tax, and reporting requirements of its markets (product pages, site profile).

From March to August 2024 I worked on it as a Python Programmer, contributing to the platform’s business workflows as part of HashMicro’s product engineering team.

The interesting engineering problem in that work is not any individual transaction. It is the fact that one business event has to be consistent across several functions at once: a confirmed order moves stock, creates a receivable, affects cost of goods, and appears in reporting. The platform is the integration, and the integration is the product.


Engineering Challenge

An ERP platform sits under the operational records of a business. That produces a different set of constraints from a single-purpose application:

  • one shared data model is consumed by many modules, so a change in one place is a change everywhere;
  • business workflows span modules rather than living inside a page or a screen;
  • business rules change with policy, regulation, and how the company actually works;
  • historical data matters: last year’s transactions must still be explainable, not just today’s;
  • accounting and tax output has to satisfy external authorities, not only internal users;
  • configuration has to absorb variation without forking the product per customer;
  • deployments are long-lived, so compatibility and migration paths matter as much as new features.

The practical consequence is that “adding a feature” in this domain usually means changing the behaviour of a workflow that other modules already depend on.


My Role & Contribution

My recorded contribution is Python development on the platform’s business workflows, in a product whose scope covers finance, sales, inventory, procurement, and HR operations.

That work sits inside a modular business application with a shared data model: business objects, their relationships, the rules that govern their transitions, and the consequences those transitions have in other modules. Public third-party documentation of HashMicro product development describes this stack as an Odoo-framework-based application developed in Python and XML (public documentation), which is consistent with the Python role.

Within that shape, the engineering concerns are:

  • implementing business logic against an existing domain model rather than a greenfield schema;
  • making a workflow’s effects visible in every module that depends on it;
  • respecting module boundaries so that one area’s requirements do not leak into another’s;
  • keeping behaviour compatible with data that already exists;
  • distinguishing what belongs in configuration from what genuinely requires code;
  • working in a shared codebase where another engineer’s module is a dependency, not a detail.

I am deliberately not assigning myself module ownership, customer outcomes or performance numbers here: the record of this engagement is a role and a product scope, and I would rather leave a claim out than inflate it.


Platform Architecture

Conceptually, the platform follows the common ERP shape: several business applications on top of one shared data model, with a web client and integration boundaries around them.

Finance, sales, inventory, procurement, and HR sit on a shared business data model that persists to a transactional store and integrates with tax, banking, marketplace, and logistics services.

Mermaid preview

The modules are separate products from a user’s perspective and one system from the data’s perspective. Most of the difficulty in the domain lives in that gap.


Business Workflow Path

A single business event fans out across the platform. A sales flow is the clearest example.

A quotation becomes a sales order, reserves stock, is delivered and invoiced, then becomes a receivable that posts to the general ledger and appears in financial reporting.

Mermaid preview

Each arrow is a module boundary, and each is a place where an assumption from one module meets the requirements of another. Building the flow so that it stays correct under partial completion, corrections, and reversals is the recurring problem.


Engineering Principles

The shared model is the product

A feature is valuable to the extent that the rest of the system can rely on it. Behaviour that satisfies one screen but contradicts what another module records is a defect, even when it looks correct in isolation.

Modules are boundaries, not folders

Module separation only means something if data ownership and dependency direction are respected. Where that breaks down, every subsequent change becomes a coordination problem.

Configuration before customization

Variation between businesses should be absorbed by configuration where the domain allows it. Code-level customization increases the cost of every future upgrade, so it has to earn its place.

Business data is long-lived

Transactional records outlive the code that produced them. Compatibility is a requirement, not a cleanup task, because a migrated record has to remain interpretable.

Local requirements are part of the domain

Accounting, tax, and reporting rules are external contracts. Treating them as configuration details rather than domain requirements produces software that is technically correct and operationally unusable.


Public Product Context

According to HashMicro’s public materials and coverage, the platform provides:

  • modular business applications covering finance and accounting, sales and CRM, inventory and warehouse management, procurement, manufacturing, HR and payroll, and point of sale;
  • cloud deployment with an on-premise option;
  • localization for regional financial and tax regimes, including Indonesia, Singapore, Malaysia and the Philippines;
  • integrations with marketplaces, banking, and logistics providers;
  • an AI assistant, Hashy, introduced in 2025, and cloud AI infrastructure work with Alibaba Cloud.

The vendor states that the platform is used by more than 3,000 businesses. Features such as the AI assistant and the Alibaba Cloud work post-date my engagement and are cited here only as product context, not as my contribution.

Public References

Scope, Limits & Confidentiality

HashMicro ERP is a proprietary commercial product and this was an employment engagement, not a personal project.

This case study therefore states:

  • product capabilities only as they are publicly documented;
  • my involvement only at the level that is recorded - a Python programming role on the platform’s business workflows - and does not attribute module ownership, internal benchmarks, customer outcomes or metrics to me.

It intentionally omits:

  • proprietary source code and internal implementation details;
  • internal architecture and infrastructure topology;
  • customer data and confidential business rules;
  • internal processes, roadmaps, and team structure.

The diagrams above are conceptual representations of the ERP engineering domain. They are not reproductions of any internal architecture.


Why This Work Matters

Most of what I work on now - microservices, distributed systems, retrieval pipelines - is about coordinating independent components that must agree on shared state.

ERP is where that problem exists in its most uncompromising form. The components are business functions with different owners, different priorities and different notions of when a record is “done”, and the state they have to agree on is the company’s financial reality. There is no grace period for inconsistency, because the ledger is the system of record.

Working on that platform is why data ownership, workflow correctness and upgrade cost are the first things I look for in a system rather than refinements to add later.

Search this site