> For the complete documentation index, see [llms.txt](https://jaswanthreddy.gitbook.io/home/llms.txt). Markdown versions of documentation pages are available by appending `.md` to page URLs; this page is available as [Markdown](https://jaswanthreddy.gitbook.io/home/home.md).

# home

I’m Jaswanth Reddy, a computer vision and NLP researcher with a BTech from IIT Tirupati. I grew up riding oxcart through the dusty lanes of my village in India, a small memory that still grounds me in where my journey began

My focus is on building practical, reliable ML models. I’ve worked across 3D vision, tracking, and language technologies.

I began my research journey at [VAL(IISc)](https://val.cds.iisc.ac.in/), India’s top-ranked computer vision group on [CSRankings 2024](https://csrankings.org/#/fromyear/2014/toyear/2024/index?ai\&vision\&mlmining\&nlp\&inforet\&in). There, I worked on 3D reconstruction and object manipulation using Neural Radiance Fields and Gaussian Splatting.

After IISc, I joined [ViSAL Lab](https://sites.google.com/iittp.ac.in/ee-spcv-lab?usp=sharing), where I develop a memory-based long–short term tracking model for transformer architectures. My work spanned both single and multi-object tracking.

I’m currently at [AutoAlign.ai](https://www.autoalign.ai/), building next-generation NLP and GenAI systems focused on trustworthiness, safety, and production reliability. My work includes: \
• Designing LLM guardrails for sensitive and high-stakes workflows\
• Building RAG and Agentic pipelines, evaluation systems, and scalable data annotation frameworks\
• Developing response modeling, code extraction, and intelligent document understanding systems\
• Creating end-to-end ML pipelines for customer intelligence, automation, and decision systems

Across projects, I’ve worked on LLM orchestration, Agentic workflows, vector search systems, predictive modeling, and AI infrastructure that bridges research with deployment.

If you’re building in computer vision, NLP, GenAI, or applied ML systems, I’d love to connect.

### Latest Updates

***

* <mark style="color:$info;">DEC 1,2025</mark>              [Reduce and Refine](https://tarka-air.gitbook.io/home/tarka-v1/reduce-and-refine-6-equal-to-28)
* <mark style="color:$info;">NOV 11,2025</mark>            [Tarka Embedding V1](https://tarka-air.gitbook.io/home/tarka-v1/tarka-embedding-v1)
* <mark style="color:$info;">NOV 8,2025</mark>             [UniC-Lift accepted at AAAI 2026](https://unic-lift.github.io/)
* <mark style="color:$info;">MAY 11,2025</mark>           [Tiny Url](/home/blogs/tiny-url.md)
* <mark style="color:$info;">MAY 11,2025</mark>           [Factcheck in SLM](/home/blogs/factchecking-in-slm.md)

<p align="center"></p>

<p align="center"><a href="https://huggingface.co/Jaswanth-0821" class="button secondary" data-icon="face-smiling-hands">Hugging Face</a>   <a href="https://www.linkedin.com/in/jaswanth-reddy-p/" class="button secondary" data-icon="linkedin-in">LinkedIn</a>   <a href="https://x.com/Jaswanth_034?t=jqxTx_o2FRjCwyp8_iUDFQ&#x26;s=09" class="button secondary" data-icon="x-twitter">Twitter</a>  <a href="https://github.com/jaswanth-0821" class="button secondary" data-icon="github">Github</a></p>

<p align="center">Reach out to me</p>