I’m Jaswanth Reddy, a BTech graduate from IIT Tirupati, specializing in Computer Vision and NLP. I worked as Research Intern in VAL Labs (IISc), which is ranked 1st in Computer Vision in India according to CSRankings 2024. There, I contributed to cutting-edge projects focused on 3D reconstruction and manipulation using Neural Radiance Fields and Gaussian Splatting. After that, I worked with ViSAL Labs, where I developed a memory-based long and short-term tracking model for transformer-based architectures, specializing in single and multi-object tracking.

Currently, at Samasra Soft, my primary focus is on advancing NLP technologies. I’ve played a pivotal role in developing robust safety guardrails for large language models (LLMs) and led the creation of comprehensive datasets and pipelines for efficient training and testing.

My primary career objective is to deepen and solidify my expertise in multi-modal Large Language Models (LLMs) and reinforcement learning (RL). I aim to leverage this knowledge to excel as a machine learning engineer specializing in LLM architecture development and LLMOps. To achieve this, I have crafted a clear, strategic roadmap that targets key areas of study and practical application, ensuring I remain at the forefront of NLP and AI innovations and become a recognized expert in the field.

👇🏻 Technologies I have had contact with or have closer contact with in my daily work and am learning today. I have a bit more familiarity and am in a continuous learning process with these technologies.

Python	Git	GitHub	Jupyter	Docker	Pytorch	FAST API	HuggingFace	Tensorflow
Google Cloud	AWS	BitBucket

👇🏻 The main technologies I intend to develop, enhance, or study in the future.Technology I have little familiarity with at the moment but have a lot of interest in learning.These technologies/frameworks are definitely on my list for future studies..

React

Typescript

Terraform

Jenkins

Kubernetes

👇🏻 Here are some projects I've been working on—whether they are early learning projects, ones I developed after gaining more knowledge, or recent ventures. Get to know some of them:

Status	Project Name	Description
In Progress	BRAG	I am currently developing a unified Multi-Modal RAG framework that integrates various applications and research papers. This framework combines retrieval-based and generative models to enhance the generation of accurate, context-aware responses by leveraging external knowledge sources. It’s still in its beta stage, but I’m excited to share it soon for everyone to use.