Skip to main content
Kunal Bhusare
Release notesCurrent build v6.0

I make
small teamsship like
big ones.

Chief Product Officer at TEX. Over nineteen years I’ve been the engineer, the agency founder, the product lead and the founder. Now I help founders and product teams rebuild how they ship — around AI.

AI Build SessionsRead the release notes
Kunal Bhusare, smiling, in glasses and a linen blazer, seated on a grey sofa
Fig. 01The author, in office hours

  1. 01

    19

    years building software, from engineer to CPO

    Patni, 2007 → TEX, today

  2. 02

    2→35

    people at Gyrix, bootstrapped, in three years

    Gyrix TechnoLabs, co-founded 2012

  3. 03

    2–3 days

    per feature, down from 1–2 weeks

    TEX for Schools, AI-first workflow

  4. 04

    >50%

    fewer bugs from new releases

    TEX for Schools, AI pre-review

§01Changelog

Release notes.

Nineteen years, seven releases, a few breaking changes. Newest first — like any good changelog.

Version 6.0

2024 — now

CurrentBreaking change: AI-first

AI-first product & engineering

TEX for Schools · AI Build Sessions Sydney

Rebuilt how the TEX for Schools team ships, with AI as the operating model instead of a helper: agentic IDEs, background agents for scaffolding, tests and review, and prompt playbooks as institutional memory. Now I help other teams do the same.

Shipped

  • Feature delivery cut from 1–2 weeks to 2–3 days
  • Code review turnaround from 1–2 days to same-day, via AI pre-review
  • Bugs from new releases down by more than half
  • New-developer onboarding from ~2 weeks to under 5 days
  • Published the Agentic Development Ladder, a free field guide for AI-first teams
  • Added a leadership edition of the ladder self-assessment, and a check-by-check comparison with engineering
Learned

AI doesn't speed up the old process. It rewards teams willing to change the process.

Version 5.0

2023

PausedExperimental: Kiki

Founder

Kiki Prep Sydney

Built Kiki — a bespectacled koala — an AI study companion that helps NSW students prepare for the Selective test and the HSC through short conversations, micro-learning and spaced repetition.

Shipped

  • A conversational AI tutor for Selective and HSC preparation
  • Learning loops built on micro-lessons and spaced repetition
  • Featured in Western Sydney University's critical-thinking video series
Kiki, a koala in glasses and a tweed waistcoatPaused
Learned

The hard part of learning isn't the content. It's coming back tomorrow.

Version 4.0

2019 — now

CurrentMajor: Product

Product Manager → Chief Product Officer

TEX Sydney

Joined TEX, a video-first marketplace that connects people one-to-one with the right expert at the right time. Turned founder vision into roadmap and platform, then led TEX for Schools — SaaS that helps Australian school counsellors connect students with mentors, pathways and alumni.

Shipped

  • Product strategy and roadmap, from founder vision to a live marketplace
  • TEX for Schools, for counsellors, students and alumni
  • Planning and delivery changes that lifted developer productivity by 20%
  • SAFe Scrum Master (2019) · AWS Certified Cloud Practitioner (2021)
Learned

A roadmap is a list of bets. Say which ones you'd lose first.

Version 3.0

2017 — 2019

ShippedDeploy: Sydney

CEO & Onsite Delivery Manager

Gyrix TechnoLabs Sydney

Moved onsite to Sydney to run the Australian side of the business — client engagement, scoping, estimation and planning — with the engineering team offshore. Shipped e-commerce and marketplace builds on WordPress, Magento, Shopify, React and Node.

Shipped

  • End-to-end delivery for Australian e-commerce and marketplace clients
  • Agile delivery that landed projects on time and on budget
  • QA and testing processes across concurrent projects
Learned

Most projects are late at the estimate, not at the code.

Version 2.0

2012 — 2017

ShippedFork: Gyrix

Co-founder & Project Manager

Gyrix TechnoLabs India

Co-founded a development agency for startups, e-commerce and marketplaces. Ran delivery across several small teams at once — and wrote a lot of the code myself.

Shipped

  • B2B and B2C products for a wide range of clients
  • The Agile delivery framework the whole company ran on
  • Awarded 'Best App Development Company' in Central India

2 → 35

people in three years, bootstrapped

Learned

Hiring is the product, when you're the agency.

Version 1.5

2011 — 2013

ShippedPatch: Business logic

MBA (MMS), Marketing

SIMSREE, University of Mumbai Mumbai

Went back to school for the half of the job engineers don't see: markets, money, and why people actually buy.

Version 1.0

2007 — 2010

ShippedHello, world

Software Engineer

Patni Computer Systems (now Capgemini) India

Started as an engineer serving Fortune 500 clients: building and maintaining enterprise applications, working the ticket queue, and learning what 'production' really means. Before that: a B.E. in Computer Science from the University of Pune.

Shipped

  • Enterprise applications on Lotus Notes
  • A Six Sigma project to make delivery more efficient
  • 'Pat on the Back' award for on-time delivery on a new technology

§02Selected work

From sprint-bound to same-day.

TEX for Schools had a lean team, three user roles and a sector that doesn’t wait. Instead of hiring, we changed how the team builds.

Case study · 8 min read

Driving workflow transformation through AI-first development

The goal: turn engineers into designers of intent rather than manual executors of code. Junior developers shipped production-grade work with confidence; senior time went back to product thinking.

  • Agentic IDE as the default
  • Prompt-driven architecture
  • Background agents for tests & review
  • Playbooks as institutional memory
Read the case study
tex-for-schools/workflow.diff+5 −5
Before and after adopting an AI-first workflow
MetricBeforeAfter
Feature delivery time1-2 weeks per feature2-3 days per feature
Code review turnaround1-2 daysSame day (via AI pre-review)
Bugs from new feature releasesModerate (20% of releases)Reduced by >50%
Team dependency on senior devHighSignificantly lower
Onboarding time for new devs~2 weeks<5 days

§03Field guide

Agents multiply your process. Including the bad parts.

The Agentic Development Ladder is the playbook I use with teams: six stages from autocomplete to a self-improving system, each with an exit checklist and the evidence that closes it. Free to read.

  1. 0AssistedHuman writes with autocomplete. Reads everything.
  2. 1DelegatedAgent writes. Human reads every line.
  3. 2HarnessedRules steer, checks catch, loops fix to green.
  4. 3Verified autonomyAgent proves behaviour. Human reads evidence.
  5. TEX is here4Managed agent teamsAgents on a queue. Human is a manager.
  6. 5Self-improving systemThe system proves its own rule changes.

One check from the self-assessment

Branch protection blocks red merges with no admin bypass, and has done so on a real PR.How sure are you?
Find your stageRead the ladder

§04Peer review

Reviewed by the people who shipped with me.

Engineers, co-founders, partners and the people I’ve managed — from LinkedIn, unedited.

It’s a pleasure to write this recommendation for Kunal, someone I have the utmost respect for. As a serial entrepreneur, Kunal has an incredible ability to come up with unique, out-of-the-box ideas, and what sets him apart is that he doesn’t just sit on them. He takes action, builds MVPs, tests them with the market, and engages with potential clients. It’s inspiring to watch him bring ideas to life and constantly refine them.

Sandesh Suradkar

Sales Application Engineer, ABB

LinkedIn ↗

§05Elsewhere

Off the roadmap.

Press, classrooms, stages, a YouTube channel about local founders — and, yes, an IMDb credit.

Off the clockSES flood rescue volunteer · Red Cross blood donor · Clean Up Australia · Mentor to engineering students.

§06Now booking

AI Build Sessions.

Hands-on sessions for founders and product teams who want AI to change how they ship — not just how fast they type. We work on your codebase: tooling, agents, review loop and the metrics to prove it worked.

  1. 01

    Diagnose

    Map where delivery actually slows down — planning, coding, review, QA or release.

  2. 02

    Rewire

    Set up AI coding agents, project rules and playbooks on your real codebase.

  3. 03

    Measure

    Agree benchmark metrics and a follow-up plan so the gains stick.

Book a sessionEngagements and pricing

§07Contact

Let’s ship
something.

Working on something where a small team needs to move like a big one? Tell me what you’re building.

Send an emailLinkedIn

Sydney, AustraliaLocal time