Vishesh Tripathi Email & Phone Number

Faridabad, Hr, In

Bengaluru, Karnataka, India

• Worked on the first Indonesian Large Language model Komodo-7B, which tends to outperform the existing models like GPT-3.5, Llama-2, Mixtral. Was involved in data preparation and preprocessing, benchmarking the models on the existing and new evaluation benchmarks.• Designing and executing experiments with DeepSpeed and VLLM for Inference aimed at optimising model inference speeds.• Worked on debugging the pruned Llama-3 model and specifying more evaluation metrics to measure the capability of the model.

Bengaluru, Karnataka, India

• Leading advancements in data parsing and extraction, emphasising proficiency with both complex and simple tables. Established a performance benchmark for evaluation, achieving an impressive 12% improvement in answering table queries through innovative methodologies.• Leading initiatives to elevate model summarisation through strategic knowledge list reranking within the RAG Architecture framework. Conducting comprehensive assessments of various open source and closed source reranking models to optimize model performance and enhance summarisation capabilities.• Successfully mitigated GPT-3, Elastic Search call issues by developing new test cases, ensuring seamless integration and performance robustness.

• I worked on a project titled REM Sleep Classification using Ensemble Learning and Score Level fusion. • As part of the project, I developed a methodological framework to build a model to differentiate between REM and Non-REM behavioral disorders. • I was also awarded the best project award and a cash prize based on my performance. • Furthermore, I trained and separated models including Naive Bayes, and Gradient Boosted trees, then combined the metrics of all models to put in comparison and visualization. After carefully observing the results and visualization of the comparison, I performed a score level fusion for the same.

India

• Built the product from scratch and constructed 2-3 roadmaps of technology followed throughout the timeline • Won competitions like Microsoft Imagine Cup, FIIB's Meraki, and SIH 2022.• Worked on technologies like .NET, ML Studio, C-sharp, HTML, Microsoft Azure, IoT.• Also worked on the growth of products by contributing to the Revenue Model and go-to-market strategy.

• As a Junior Machine Learning Engineer, participated in 2 community projects including Road Safety in India and Racism in news media around the world.• Used and learned techniques like Web Scrapping, Data collection, Data cleaning and Algorithm Build-up.• Contributed to the community project with a team of 24 members and learned a lot of things.

Bachelor Of Technology - Btech, Computer Science With Specialization In Ai & Ml (Bhopal Campus), 8.1

Higher Secondary, Computer Science, 88%

Activities and Societies: Aryabhatt House Captain (Winning House Captain) 2019-2020

Matriculation, 81%