Skip to content
Bijak (Krishiacharya Technologies)

Bijak

Putting a trust-based, cash-heavy trade into an app — in eleven languages.

The tension

Mandi trade runs on relationships and cash. The app had to earn the trust of traders who’d never used one — while money moved through nearly every screen.
Role
Engineering Lead · Jul 2020 — Oct 2024
What I owned
Involved from day one — led the move from native to Flutter, set the architecture and release process, and led a team of 8 engineers shipping alongside product, design and marketing.
Platforms
Android · iOS
downloads, incl. Just Fresh
600K+
languages
11
mandis with daily prices
2,000+
  • Bijak — buy commodities from 30,000+ verified traders
  • Daily mandi prices across 2,000+ mandis and 200+ commodities
  • Language picker with 11 Indian languages
  • Payment guarantee and Bijak Limit credit facility
  • Trader ratings to pick trustworthy buyers and suppliers
  • Digital bookkeeping with all transactions in one place
Bijak website — Optimization through Accountability

Context

Bijak is a B2B agri-trade platform for India’s mandi traders. Buyers and suppliers use it to find 30,000+ verified traders, check daily prices from 2,000+ mandis, and trade with services like advance payments, a payment guarantee and the Bijak Limit credit facility — plus digital bookkeeping for every transaction.

I was involved from the beginning — leading the app’s migration from native Android to Flutter, growing into Engineering Lead of a team of 8, and later also owning the Just Fresh apps.

Engineering nuance

1. Migrating to Flutter without slowing the business down

Constraint. We were moving from a native app to Flutter while the business kept asking for features. The rewrite couldn’t become a freeze.

Decision. Flutter 3 with GetX and MVVM — at the time, GetX was the natural choice for building fast: state, routing and dependency injection in one lightweight package, with very little boilerplate. One way of doing things across the app (and later across Just Fresh).

Trade-off. GetX is opinionated and less “pure” than some alternatives, but it let a team of 8 migrate and keep shipping at the same time — which mattered more than architectural elegance.

2. Language is a trust feature

Constraint. Many traders were first-time smartphone users. An English-first app would have been an app they didn’t trust.

Decision. Make localisation first-class: 11 languages, with a language picker up front and layouts that hold up when strings get long.

Trade-off. Every feature costs more to ship and test — but adoption depends on it.

3. Measure every funnel, automate every release

Constraint. Marketing, product and operations all needed answers — which campaigns brought traders in, where they dropped off — and releases couldn’t be a bottleneck.

Decision. Segment, MoEngage and Branch.io for analytics, engagement and attribution; Codemagic CI/CD to automate builds, tests and store releases.

Trade-off. More tooling to maintain, but decisions moved from opinion to data, and shipping became routine.

4. AI where it removes manual work

Behind the app, I trained TensorFlow models and introduced Gen AI tooling to cut human intervention in internal processes — quiet automation that saved the operations team time.

Leadership

  • Led a team of 8 engineers, coordinating with design, product and marketing on a predictable cadence.
  • Ran code reviews and mentored developers on architecture and Flutter practice.
  • Built B2B and D2C products on a shared foundation.

Outcome

  • 600K+ downloads across Bijak’s apps, including Just Fresh.
  • A Flutter foundation that carried over to the Just Fresh apps.

Next case study

Engineering LeadJul 2022 — Oct 2024

Just Fresh

A family of apps on one codebase — 80% of it shared — for produce that changes price every morning.

downloads (within Bijak’s 600K+)
50K+
code shared across the apps
80%
Read case study
Just Fresh website — Just Fresh. All Nature. No Nasty.