badal.sahani

Badal Sahani · Senior Frappe/ERPNext Engineer · Tech Lead

Every transaction posted, verified, reconciled.

I build revenue-grade platforms — ERP and payments at the core, with the AI tooling, realtime systems, and cloud infrastructure around them — and lead the 8–10 engineers who ship it all. Hands-on daily, from schema to Terraform to the LLM pipeline reviewing our PRs.

Journal · FY 2023 – present

  1. Fee transactions verified, deduplicated, reconciled

    HMAC-signed webhooks, idempotent, replay-safe

    100%
  2. Custom Frappe apps in production

    thin overlays over one shared core — no forks

    25+
  3. Users served across institutions

    4–5 education groups, one multi-tenant core

    20–30k users
  4. Payment gateways integrated

    behind one pluggable router, HMAC-verified

    6 gateways
  5. Bank payout rails connected

    incl. ISO 20022 over GPG-encrypted SFTP

    8 banks
  6. Notices API latency, before → after

    query restructuring + Redis caching

    4–5s → <500ms
  7. Infrastructure cost reduction

    documented, with a raised security posture

    −43%

Selected work

all work →

Fees & Finance Engine

Payments

The platform's entire revenue system — 40+ DocTypes on ERPNext, from payment plan to GL posting.

engines
payment plans · instalments · deferrals
ledger
discount stack posting GL entries
entries
multi-company split Payment Entries
migration
v2 re-architecture, zero downtime

An EMI lifecycle state machine with single-writer settlement eliminated double-booked payments; auto-posted settlement Journal Entries removed manual reconciliation.

ERPNext · Python · MariaDB · GL / Journal Entries

LLM Review Pipeline

AI Engineering

An automated code-review pipeline: every pull request is read by an LLM, grounded in the changed code's context, before a human ever sees it.

pipeline
diff + repo context → LLM → PR review
grounding
review prompts built from the changed code
gates
runs beside Semgrep + dependency audits
adoption
every PR, 10-person team, daily

Reviewers start from a machine-annotated diff instead of a cold one — cutting review turnaround and catching the boring bugs before humans spend attention on them.

LLM APIs · CI/CD · GitHub · Python

Multi-Tenant ERP Core

System Design

One shared core serving 4–5 institutions and 20–30k users — 25+ custom Frappe apps, 250+ DocTypes.

tenancy
institution as a first-class dimension
security
row-level permission_query_conditions
overlay
thin per-institution apps over hooks
upgrades
no forks, no monkey-patching

New institutions onboard without forking the platform; the core stays clean through Frappe version upgrades.

Frappe · ERPNext · Python · 500+ REST endpoints

Four disciplines

how I work →

System Design

Multi-tenant architecture where correctness is the feature: one core, many institutions, no forks.

AI Engineering

AI as infrastructure, not garnish — wired into the team's daily loop where it compounds.

DevOps & Platform

Infrastructure as code, delivery as GitOps, secrets as references — nothing hand-built, nothing plaintext.

Payments Engineering

High-volume fee collections through code I own — every transaction posted, verified, and reconcilable.

Working stack

  • Frappe
  • Python
  • TypeScript
  • React
  • Redis
  • MariaDB
  • Socket.IO
  • Claude
  • Copilot
  • AWS
  • Kubernetes
  • Terraform
  • Helm
  • ArgoCD
  • Docker
  • Nginx
  • Git
  • Bash

Building on Frappe and need someone who has shipped it at scale — payments, upgrades, infrastructure, team and all?