Skip to content
Vishal
Vishal Goyal
  • Software engineering
  • AI / ML
  • Open source

I build software and intelligent systems.

Python developer exploring AI/ML, APIs, AI agents and real-world engineering systems. B.Tech CSE (AI & ML) student, learning by building — and inside real codebases.

Explore my workLet’s connect
system.graphactive
DATAMODELTOOLSAGENTPRODUCTFEEDBACK
DATA
collect, clean, represent
MODEL
inference, evaluation, baselines
TOOLS
retrieval, functions, io
AGENT
reason, act, observe
PRODUCT
interface, feedback, iteration
nodes 05edges 05
System status
Building
Focus
Software engineeringAI / MLOpen source
Studying
B.Tech CSE (AI & ML)ITM University · 2025–2029
Open to
Software · AI/ML workOpen-source collaboration
01About / Now

Software first. AI/ML is the direction.

I’m a B.Tech Computer Science (AI & ML) student, and almost everything I know came from building things and reading other people’s code.

Software engineering is the foundation — Python applications, APIs, databases, the parts that simply have to work. AI/ML is the direction I’m building toward, not a title I’ve earned yet.

Open source is where that gets tested: unfamiliar codebases, real bugs, and maintainers who review what I write.

  • 01

    Build to understand

    Reading about a system teaches you the vocabulary. Building it teaches you the system.

  • 02

    Baselines first

    Cosine similarity before a fine-tune. Postgres before a vector store.

  • 03

    Read the source

    The answer is usually already in someone else’s repository.

Currentlylive

$ Building

  • Python applications
  • APIs & backend systems
  • AI systems
  • Web software

$ Interested in

  • Software engineering
  • AI / ML
  • Intelligent agents
  • Debugging
  • Testing

$ Exploring

  • Open source
  • Linux
  • Cloud-native tooling

$ Build log

understandexplorebuilddebugiterate

02Selected work

Things I’ve built, honestly labelled.

Projects

03

Ongoing

01

03Open source

I learn by working in real codebases.

Contributing across AI/ML infrastructure, climate technology, developer tooling and software systems — reading unfamiliar code, debugging real problems, and getting changes through review alongside maintainers.

18

Merged pull requests

across

  • AI / ML
  • Climate tech
  • Developer infrastructure
  • Software systems

Ecosystems I’ve contributed to

  • Climate technology

    Open Climate Fix

    • Solar Consumer
    • Quartz Solar Forecast
    • Graph Weather
    • Database support & tests
    • Country-code handling
    • Retry logic
    • Monitoring
    • Forecast fixes
    • CI & workflow improvements
  • AI / ML infrastructure

    Kubeflow / Kale

    • KFP securityContext error handling
    • Node.js 22 CI workflows
    • Disabled-GPU dialog dark mode
  • ML infrastructure

    PyTorch Lightning

    • Checkpoint loading
    • weights_only
    • Hyperparameter handling
    • LightningModule
  • Developer infrastructure

    Microsoft AzureTRE

    • CI workflows
    • Security checks
    • pytest-asyncio
    • GitHub Actions
  • Web infrastructure

    Django

    • Routing & redirects
    • Template internationalization
    • Docs
  • Learning resources

    EbookFoundation

    • Link maintenance
    • Broken-link fixes
  • AI / developer infrastructure

    Langflow / OpenRAG

    • Documentation alignment
    • OpenSearch index initialization
    • kNN configuration
  • Generative agent research

    Google DeepMind / Concordia

    • Memory-state deserialization
    • Malformed choice-action specifications
    • Python development tooling
  • Python tooling

    Google etils

    • Duplicate logging behaviour

Merged contributions

A handful of examples, not the archive — the rest is on GitHub.

Additional open-source work

  • Generative agent researchClosed

    Google DeepMind / Concordia

    • Memory-state deserialization
    • Malformed choice-action specifications
    • Python development tooling
  • Python toolingOpen

    Google etils

    google/etils#765
    • Duplicate logging behaviour
  • AI / developer infrastructureOpen

    Langflow / OpenRAG

    langflow-ai/openrag#847
    • OpenSearch index initialization
    • kNN configuration

These particular contributions are still open, or were closed without being merged. They are deliberately excluded from the count above.

No contribution graphs here — the full history is on GitHub, where it can speak for itself.

View GitHub

How I build

  1. 01

    Understand

    Start with the problem. Most of the work is deciding what actually needs to exist.

  2. 02

    Explore

    Read the documentation, the code and the research until the shape of it is clear.

  3. 03

    Build

    Turn ideas into working software — typed, inspectable and runnable by someone else.

  4. 04

    Iterate

    Test, debug and improve. Then go again with whatever the last pass taught me.

04Technical stack

What I actually reach for.

Not a logo wall. Grouped by where each thing sits.

Python first. The rest follow the problem.reach for first
Python
C
JavaScript
TypeScript
HTML
CSS
05Experience / Education

Where the fundamentals are coming from.

Structured programmes and coursework alongside the building — labelled for exactly what they are.

  • Job simulationAug 2026

    GenAI Powered Data Analytics

    Tata iQ · Forage

    Analytics work with generative AI in the loop — from exploring the data to reasoning about how a model would be used responsibly.

    What it covered
    • AI-powered data analytics
    • Exploratory data analysis
    • Data quality assessment
    • Risk indicators
    • Predictive modelling framework
    • Customer delinquency risk
    • GenAI-assisted model logic & evaluation
    • Agentic AI collections strategy
    • Ethical AI
    • Regulatory compliance
  • Job simulationAug 2026

    Data Visualisation

    Tata Consultancy Services · Forage

    Turning raw retail transaction data into figures a business could actually act on.

    What it covered
    • Data cleaning
    • Visualisation
    • Business insights
    • Retail transaction analysis
    • Revenue calculations
    • Monthly revenue
    • Top countries & customers
    • Product demand

Forage job simulations are self-directed programmes that mirror real client work. They are not employment, and they are not listed here as such.

Education

ITM University

2025 — 2029

B.Tech in Computer Science — Artificial Intelligence & Machine Learning

Gwalior, India

Selected

Center of Excellence

Selected from Semester 2

  • Software
  • Data
  • AI
06Contact

Let’s build something interesting.

Open to interesting software, AI/ML, open-source and engineering opportunities. If you’re building something I’d learn from, I’d like to hear about it.

Start a conversation

I read everything. I’m slower to reply when I’m deep in a build.