# ZeroShotLab — AI engineering and digital experiences

> ZeroShotLab is an independent AI engineering, web design and Android development studio led by Randhir Kumar in India.

## What ZeroShotLab does

- LLM fine-tuning: data preparation, adapting open language models, evaluation and local model formats.
- Retrieval-augmented generation (RAG): document question answering, retrieval, passage verification and source citations.
- AI agents and automation: assistants, tool-connected workflows and integrations with human handoffs.
- Web design and development: websites and interfaces that connect AI engineering with a usable product.

- Android app development: interfaces, API integrations, useful AI features, testing and release preparation.

## Selected engineering work

- [HumanFlow](https://huggingface.co/randhir302/HumanFlow): a fine-tuned Llama 3 8B model for natural prose, with GGUF versions for local use.
- [Corrective RAG](https://github.com/iamhero2709/corrective-rag): document Q&A with passage verification, citations and abstention when evidence is missing; Python, NLI, FastAPI and Docker.
- [MoEVL-Tiny 0.5B](https://huggingface.co/randhir302/MoEVL-Tiny-0.5B): compact vision-language model with SigLIP, Qwen2.5-0.5B and mixture-of-experts layers.

## Research

- [Verification Without Sufficiency](https://arxiv.org/abs/2608.00585): research on per-chunk filtering and multi-hop RAG.
- [Holographic Memory for Zero-Shot Compositional Reasoning](https://arxiv.org/abs/2606.24948): mechanistic study of compositional reasoning in knowledge graphs.
- [Trail-Weighted Influence Scores](https://arxiv.org/abs/2610.27366): approximate outcome prediction in Bayesian networks.

## Website concept portfolio

The website demos are concept work, not client commissions. Businesses shown are not affiliated with ZeroShotLab.

## About the engineer

Randhir Kumar is an AI engineer in India. Public profiles:

- [GitHub](https://github.com/iamhero2709)
- [Hugging Face](https://huggingface.co/randhir302)
- [LinkedIn](https://www.linkedin.com/in/randhir-kumar-861573301)

## Starting a project

Email randhir2709vns@gmail.com or use the enquiry form at https://zeroshotlab.in/#contact. Share your goal, relevant links, current stack and constraints. Scope and timeline are discussed before work is agreed. Public fixed prices are not listed.

## Common questions

### What does ZeroShotLab do?

ZeroShotLab is Randhir Kumar’s independent AI engineering, web design and Android development studio in India. Work includes LLM fine-tuning, retrieval-augmented generation (RAG), AI agents, model evaluation, web interfaces and Android app development. Projects start with a conversation about your goal and current setup.

### Should my project use RAG or LLM fine-tuning?

RAG retrieves relevant information at answer time and is useful when answers need current or private documents and source references. Fine-tuning changes a model’s behavior using training examples and can help with a specific task, style or output format. A project may use either or both; the choice depends on the data and evaluation goals.

### What AI engineering work can I review?

HumanFlow is a fine-tuned Llama 3 8B model published on Hugging Face. Corrective RAG is an open-source document Q&A system with passage verification and source citations. MoEVL-Tiny 0.5B is a compact vision-language model. The engineering portfolio links directly to the code or model pages.

### How do you evaluate an AI system?

Evaluation should reflect the task. For a retrieval system, check whether it finds the right evidence, cites it correctly and declines unsupported answers. For a fine-tuned model, compare it with the base model on representative held-out examples. We agree on the evaluation criteria as part of the project scope.

### Can you work with my existing models or codebase?

Yes. Share your current model, data sources, codebase or API setup and what you want to improve. I can help with a focused part of an existing system or discuss a new implementation based on the scope.

### Who owns the code and project accounts?

You retain your project code and accounts as agreed in the project scope. Third-party models, datasets and dependencies keep their own licenses, which need to be checked for the intended use.

### How do we start a project?

Use the project enquiry form or email randhir2709vns@gmail.com. Share your goal, relevant links, current stack and constraints. We discuss the technical approach, scope and timeline before agreeing on the work.