Shrikara Varna

AI Researcher

Mountain View, California, United States2 yrs 4 mos experience

Most Likely To SwitchAI Enabled

Key Highlights

Expert in large language models and machine learning.
Proven track record in improving AI model performance.
Strong background in both academic and practical applications.

Stackforce AI infers this person is a Machine Learning Engineer with expertise in SaaS applications.

Contact

Skills

Core Skills

Machine LearningDeep LearningWeb DevelopmentSoftware Testing

Other Skills

OCRVLMsLLMsrule-based conversion systemsTensorFlowComputer VisionCode GenerationLarge Language Models (LLM)Amazon Web Services (AWS)Python (Programming Language)Hyperparameter TuningDeep Neural Networks (DNN)Question AnsweringNatural Language Processing (NLP)React.js

About

I’m a master’s student at Carnegie Mellon University’s Language Technologies Institute (LTI), pursuing a Master of Science in Intelligent Information Systems. I have a strong background in large language models (LLMs) and machine learning, complemented by practical experience in frameworks like TensorFlow and PyTorch. I’m actively seeking full-time opportunities in the field of machine learning starting in January/February 2025. I’m excited to bring my diverse skill set and passion for AI to contribute meaningfully to a forward-thinking company. Currently, I’m working with Prof. Emma Strubell on Scientific Entity Extraction from publications, focusing on information extraction and disambiguation from tables. I’m also collaborating with Prof. Carolyn Rose on Code Translation for Low-Resource Languages. Previously, I worked with Prof. Teruko Mitamura on multimodal summarization, where I developed models to summarize instructional videos. Beyond AI and ML, I bring expertise in Docker, ONNX, AWS, React.JS, and FastAPI, giving me the versatility to drive innovation in fast-paced environments.

Experience

2 yrs 4 mos

Total Experience

1 yr 2 mos

Average Tenure

1 yr 4 mos

Current Experience

Synopsys inc

2 roles

AI Senior Engineer

Feb 2025 – Present · 1 yr 4 mos · Sunnyvale, California, United States · On-site

Architected a hybrid multimodal document ingestion pipeline combining OCR, VLMs, LLMs, and rule-based conversion systems for enterprise technical documentation
Improved Knowledge Assistant retrieval efficiency from 71.4% to 93% through optimized document conversion, chunking, and retrieval workflows
Built scalable agentic planning and tool-routing systems achieving 98% routing accuracy across multi-tool execution pipelines
Developed workflow (“recipe”) extraction systems that convert engineering documentation into executable agent workflows
Integrated OpenCode into internal agentic infrastructure to improve tool execution robustness and multi-step reasoning

OCRVLMsLLMsrule-based conversion systemsMachine LearningDeep Learning

Machine Learning Intern

May 2024 – Aug 2024 · 3 mos · Sunnyvale, California, United States · On-site

Developed a code generation model to create testbench for VCS. Increased the code coverage by 12% and found 7 bugs in the tool
Built a LLM agent to help create functions in Python using classes described in a in-house library to build Verilog test bench for RTM

Code GenerationLarge Language Models (LLM)Amazon Web Services (AWS)TensorFlowPython (Programming Language)Hyperparameter Tuning+3

Carnegie mellon university

Graduate Teaching Assistant

Aug 2024 – Dec 2024 · 4 mos · Pittsburgh, Pennsylvania, United States · On-site

TA for the Generative AI course (10423/623) from MLD department at SCS
Created homework on image diffusion models (prompt-2-prompt and DDPM)
Lead recitation on the topics of diffusion models
Mentored four student projects

TensorFlowComputer VisionMachine Learning

Cisco

2 roles

Software Engineer

Aug 2022 – Aug 2023 · 1 yr · Bengaluru, Karnataka, India

Implemented an unsupervised ML model for Question Decomposition for complex multi-hop questions which had an accuracy of 82%
Designed and developed an active learning model to help identify intents in queries that resulted in the miss rate of a bot. this resulted in an improvement in chatbot performance by an average of 20%
Built the Analytics Dashboard by integrating Atlas Charts (from MongoDB) with React JS

Amazon Web Services (AWS)Question AnsweringNatural Language Processing (NLP)Deep Neural Networks (DNN)Machine Learning

Software Engineer Intern

Feb 2022 – Jun 2022 · 4 mos · Bangalore Urban, Karnataka, India

Worked on Cisco internal application - BotLite - to build Drag and Drop Feature using ReactJS.
Employed React Flow to achieve draggable components. Developed the GUI from scratch and improved the response time for the user’s request
Followed a Scrum-based Agile development process.

React.jsFastAPIWeb Development

Akamai technologies

SDET Intern

Jun 2021 – Jul 2021 · 1 mo · Bangalore Urban, Karnataka, India

Interned in the role of SDET in the Intelligent Edge QA team. Involved in automation of testing using the robotframework and Python language.

Python (Programming Language)Robot FrameworkSoftware Testing

Samsung research institute bangalore

Samsung PRISM Researcher

May 2021 – Nov 2021 · 6 mos · Bangalore Urban, Karnataka, India

Worked on building an application to detect words of different languages being present on the mobile screen and translating them to the base language of the phone.
Awarded 'Certificate of Excellence' by Samsung PRISM for the project.

Deep LearningNatural Language Processing (NLP)Recurrent Neural Networks (RNN)Machine TranslationMachine Learning

Hewlett packard enterprise

HPE-CTY Programmer

Mar 2021 – May 2021 · 2 mos · Bengaluru, Karnataka, India

Built a chatbot for answering Python related doubts using the Stackoverflow dataset.
Also involved building a classification model to tag the questions (input).

Natural Language Processing (NLP)Python (Programming Language)Conversational AIDeep Neural Networks (DNN)Machine Learning

Samsung research institute bangalore

Samsung PRISM Researcher

Jul 2020 – Feb 2021 · 7 mos · Bangalore Urban, Karnataka, India

Named Entity Popularity Determination using Ensemble Learning for disambiguation needed for virtual assistants(VAs).
Generation of dataset synthetically to train the models for better accuracy.
Awarded the 'Certificate of Excellence' by Samsung PRISM for the contributions to the project
Resulted in paper presentation at ICON 2020 NLP Conference.

Machine LearningNatural Language Processing (NLP)Ensemble LearningDeep Neural Networks (DNN)PandasSynthetic Data Generation