Every business runs on a mess.

← go on, drag the notes aroundAct I · The mess
Business Analyst · AI Transformation

JithendraVarma.Chamarthi, Business Analyst and AI Transformation in Adelaide

I understand the problem, design the fix, then build it. AI is my calculator.

NowICT Business Analyst, AI Transformation
BuildingTastibase · Adel AI Studio
BasedAdelaide, South Australia
Scroll to sort it
Life
Learn
Work
Build
Act II · The map

My life, as a process map.

Life
Learn
Work
Build
Act III · The evidence

Current state → future state.

Every project starts the same way: understand what's really happening, then design what should happen. AI does the heavy lifting; I decide what's worth building, and then I build it. These are real before-and-afters. Flip the switch on each one.

44,000+reviews read by the Tastibase model
711dishes ranked across 486 restaurants
Sentiment model accuracy
BERT85%
VADER50%
>95%osteoarthritis grading accuracy, published
~1,250pre-rendered pages on Cloudflare
6organisations' websites built or run
3roles grown through at one factory
Case 01 · Founder & Data Scientist · 2026

Tastibase

An IMDb for food. It ranks the dish, not the restaurant, from what real customers wrote.

  • Python pipeline with pandas, scikit-learn and Hugging Face Transformers.
  • Fine-tuned BERT over VADER, plus a Bayesian-shrinkage score blended with star ratings.
  • Next.js on Cloudflare Pages, programmatic SEO, verified in Search Console and Bing.
tastibase.com ↗PythonBERTNext.js
Case 02 · ICT Business Analyst · 2026

AI at an RTO

AI transformation and data integrity at Adelaide Culinary Institute, a hospitality and construction training organisation.

  • Training documents now come from Claude-built workflows, with registers for review and approval.
  • Wisenet and eSkilled reconciled with SQL; duplicate and misplaced enrolments resolved.
  • Gap analysis and document control against the 2025 Standards for RTOs.
ClaudeSQLWiseneteSkilled
Case 03 · Digital Business Analyst · 2023–26

RightLaserCut

From quotes and invoices to choosing the system the whole workshop runs on.

  • Mapped workflows with operations, sales and production to find gaps in job tracking.
  • Compared options, chose BuddySquare, ran UAT and wrote the SOPs.
  • KPI dashboards in Excel and Power BI for management reporting.
RequirementsUATPower BI
Case 04 · Founder · 2025–

Adel AI Studio

Practical AI automation, websites and SEO for Adelaide businesses, starting with a day on site.

  • Shadow Day: watch the business run, then deliver a proposal, hours saved and a working demo within a week.
  • Websites for Meridian Vocational College, ACI, Adelaide Construction Institute, Aim Care Services, RealVR and RightLaserCut.
  • AI coaching days that train teams to use Claude at work.
Appendix · Research & prototypes

The lab.

  • 2023 · DL

    Knee osteoarthritis grading

    VGG16 grades X-rays into five Kellgren-Lawrence classes at over 95% accuracy.

    Published · PES
  • 2025 · NLP

    Heroes & villains of Treasure Island

    spaCy NER and sentiment volatility resolve 149 mentions into 22 characters and classify their roles.

    UofA
  • 2024 · ML

    Accident-prone zones

    K-Means risk zones across five states of 2019–2022 crash data, validated with gradient boosting.

    UofA
  • 2022 · Web

    Homegrowns

    An e-commerce platform that lets mango farmers sell direct, with no middlemen.

    Published · JETIR
  • 2022 · IoT

    NFC returnable packaging

    Packaging that stores order and return details, with loyalty points for every return.

    2nd · IPR Expo
  • 2022 · OR

    Packaging collection routes

    Pincode clustering and Google OR-Tools solving the travelling salesman problem.

    Pilot
  • 2022 · Idea

    SafeguardHER

    Bluetooth beacons that push safety help to passengers as they board, readable without literacy.

    KIIT Ideathon
Act IV · The story

One last user story.

JV-001 · Backlog → ReadyPriority: High

As a hiring manager, I want a Business Analyst who understands the problem and can build the AI-powered fix, so that problems get solved, not just documented.

Acceptance criteria
  • Understands the business first: requirements, process maps, user stories, UAT and SOPs.
  • Uses AI the way an analyst uses a calculator: Claude workflows, SQL, Python and models to build the fix fast.
  • Has shipped in production: Tastibase is live, and six organisations run on sites I built.
  • Master of Data Science, University of Adelaide. Based in Adelaide; on-site, hybrid or remote.
  • Open to Business Analyst, Business Systems, Data and AI Transformation roles.

Every box ticked. That's me.

Or email[email protected]
BasedAdelaide, SA · --:--
Find me[email protected]LinkedIn
ProductsTastibaseAdel AI Studio
© 2026Back to the mess ↑