kandula.studio

About

Nikhilvarma Kandula

Founder & AI engineer, based in Germany. Also written Nikhil Varma Kandula.

Based
Germany · open to relocate
Right now
M.Sc. Big Data, FOM, to Aug 2027
Shipped
6 products live in production
Open to
Data, software and AI engineering

Hey, I’m Nikhilvarma. I build data systems, and lately I build products too.

Eighteen months at MicroIntech, a US fintech, took me from Data Engineer to Lead Developer. I rebuilt a monolithic financial platform into event-driven microservices serving 500+ concurrent users, and wired GPT-4 and Llama into an audit pipeline that cut fifteen hours of manual review a week down to about three.

Alongside it I led development at EngineeredPrompts, a premium AI-prompt platform. I owned the prompt library and the models behind it, the paid tier, and the two developers building with me.

Now I’m in Germany doing an M.Sc. in Big Data & Business Analytics at FOM Hochschule, through August 2027. The academic layer sits on top of work I was already doing rather than replacing it. Most of the analysis on this site started as a real question rather than an assignment.

Before that: a B.Tech in Computer Science from Malla Reddy College of Engineering, first class with distinction, and a peer-reviewed paper in IRJMETS on rainfall estimation that beat the standard interpolation baseline.

The best data work is invisible. The pipeline nobody notices because it never breaks. The dashboard that makes the answer obvious. That’s what I’m trying to build.

The route here

Where I have been, in order

Study, work and research ran in parallel more often than in sequence. The colour says which is which; where there is something to check, the entry links to it.

  1. Oct 2025 – Aug 2027

    StudyNow

    M.Sc. Big Data & Business Analytics

    FOM Hochschule

    In progress. Most of the analysis on this site began as a real question rather than an assignment.

  2. Mar 2025

    Research

    Published: rainfall estimation via data fusion

    IRJMETS, Vol 7 Issue 3

    Peer-reviewed. Probability of detection 0.58, beating Kriging with External Drift on the same set.

    Read the paper write-up

  3. Oct 2024 – Dec 2025

    Work

    Data Engineer → Lead Developer

    MicroIntech · US fintech

    Monolith to event-driven microservices for 500+ concurrent users at 100% transaction integrity. LLM audit pipeline cut fifteen weekly review hours to about three. Mentored five engineers.

    Read the case study

  4. 2024 – 2025

    Work

    Lead Developer

    EngineeredPrompts

    A premium AI-prompt platform. I owned the prompt library and the model orchestration behind it, the premium tier end to end, and the development team.

    Named on their team page

  5. 2020 – 2024

    Study

    B.Tech, Computer Science & Engineering

    Malla Reddy College of Engineering

    First class with distinction.

Verified elsewhere

A title someone else published

Anyone can put “Lead Developer” on their own website. This one is on somebody else’s, under their name and their team page, which is the only version of the claim worth anything.

EngineeredPrompts

Premium AI-prompt platform · founded by Gatike Shyam

Lead Developer

2024 – 2025

Where I am creditedengineeredprompts.com/about_us The platform itselfengineeredprompts.com
  • Prompt library & AI systemsThe 2,000+ prompt catalogue and the model orchestration behind it: prompts written, structured and validated as a product, not as a document.
  • The development teamCode review, architecture calls, and directing the two other developers credited on the same page.
  • Payments & subscriptionsThe premium tier end to end: paywall, checkout, and subscription state.

How I work

Five things I actually believe

Not values. Positions I have paid for at least once. Each one links to the work that taught it, including the project whose finding was that there was no finding.

  1. Define the metric before you improve it

    A percentage quoted afterwards is unfalsifiable. The 40% tech-debt reduction at MicroIntech means something only because the backlog and the remediation criteria were agreed before the work started.

    MicroIntech

  2. Know when not to trust the model

    In financial audit a confidently wrong answer is worse than no answer. The engineering is the confidence routing that sends the doubtful cases to a human. The orchestration around it is the easy part.

    MicroIntech

  3. Publish the null result

    Thirty-eight countries, eleven years, an R-squared of 0.004. A portfolio containing only confirmed hypotheses is a portfolio that has quietly deleted its failures.

    ESG and GDP growth

  4. Ship it, then write down what it cost

    There are 6 products live rather than screenshotted, and each one carries its decision record: what was chosen, over what, and what the choice gave up.

    FirstChair

  5. Plain language is a test, not a courtesy

    Every case study here opens with a paragraph a non-technical reader can follow start to finish. If I cannot write that paragraph, I have not finished understanding the problem.

    All 15 projects

Education

  • M.Sc. Big Data & Business AnalyticsFOM Hochschule, Germany · Oct 2025 – Aug 2027 · in progress
  • B.Tech, Computer Science & EngineeringMalla Reddy College of Engineering · 2020–2024 · First class with distinction

Languages

  • EnglishC1 · Professional
  • DeutschA2 → B1 · Learning
  • TeluguNative
  • HindiNative

Certifications

  • Google Data Analytics ProfessionalGoogle · Coursera · Feb 2026
  • SAP Certified Associate: Data Analyst, SAP Analytics CloudSAP · In progress
  • Big Data Engineering & ArchitectureFOM · 2026
  • Spatiotemporal Data AnalyticsFOM · 2026
  • Advanced Machine Learning for BusinessFOM · 2026
  • LLM Orchestration & Prompt EngineeringMicroIntech, on the job · 2024–25

Elsewhere

Everything under my name, in one place. The write-ups live on Medium and Substack; the code, notebooks and datasets are on GitHub and Kaggle.

Questions

Common questions

Who is Nikhilvarma Kandula?
Nikhilvarma Kandula is a founder and AI engineer based in Germany. He builds and ships AI products end to end. Ganymede, Witness, Bud, Cartwise, FirstChair and Knock are live, and he previously spent eighteen months at the US fintech MicroIntech, going from data engineer to lead developer.
What does Nikhilvarma Kandula build?
He builds AI products from the model call through to the interface. Ganymede decides which delinquent borrower is worth an agent-minute and coaches the call that follows; Witness turns a photograph of a returned gearbox part into an ISO failure record linked to its batch; FirstChair tracks how ChatGPT, Gemini and Perplexity describe a law firm against its competitors; Cartwise turns a photographed receipt into a nutrition summary; Bud is habit tracking built like a game; Knock is local services booking. He also publishes data investigations and one peer-reviewed paper.
What is Nikhilvarma Kandula’s background?
He holds a B.Tech in Computer Science from Malla Reddy College of Engineering and is completing an M.Sc. in Big Data & Business Analytics at FOM Hochschule in Germany through August 2027. At MicroIntech he rebuilt a monolithic financial platform into event-driven microservices for 500+ concurrent users and built an LLM audit pipeline that cut fifteen weekly review hours to about three.
Has Nikhilvarma Kandula led a development team?
Yes, in two places. At MicroIntech he moved from data engineer to lead developer, introduced systematic code review and mentored five engineers, and the production bug rate fell 15% over six months. At EngineeredPrompts, a premium AI-prompt platform, he was Lead Developer from 2024 to 2025 and is named in that role on the company’s own team page at engineeredprompts.com/about_us.
What did Nikhilvarma Kandula do at EngineeredPrompts?
He was Lead Developer at EngineeredPrompts, a premium AI-prompt platform at engineeredprompts.com, from 2024 to 2025. He owned the 2,000+ prompt library and the model orchestration behind it, the premium subscription tier end to end including the paywall and checkout, and the development team: code review, architecture decisions, and directing the two other developers. He is credited by name as Lead Developer on the company’s public About Us page at engineeredprompts.com/about_us.
Where is Nikhilvarma Kandula based?
He is based in Germany and open to AI, data and software engineering roles anywhere in Germany, remote, hybrid or on-site. He is also open to full-time roles in India.
Has Nikhilvarma Kandula published research?
Yes. His peer-reviewed paper on rainfall estimation by heterogeneous data fusion appeared in IRJMETS, Vol 7 Issue 3, March 2025. It fuses ground sensors, radar and satellite imagery into a single signal and reaches a probability of detection of 0.58, beating Kriging with External Drift on the same dataset.
How can you contact Nikhilvarma Kandula?
By email at kandulanikhilvarma@gmail.com, or through LinkedIn at linkedin.com/in/nikhilvarmakandula. His work and full case studies are at kandula.studio.

Open to

Engineering, data and AI roles across Germany

Available immediately, 20 hours a week on a student visa, anywhere in Germany. Data, software and AI engineering roles alike. Also open to full-time work in India.

Email meEN résuméDE Lebenslauf