Raghav Dixit Email & Phone Number

New York, Ny, Us

Building core product and infrastructure into production

San Francisco Bay Area

- Leading LanceDB Integrations with the world’s most popular tools: Building frameworks/libraries/packages for easy model loading, fast&efficient multimodal data storage, integration with lance, model training/fine-tuning and serving. Also growing existing partnerships and adding new ones.- Solving bugs for the open source version of our python library(lancedb). - Writing/Editing blogs,cookbooks, tutorials

San Francisco, California, United States

Helping build LanceDB and its APIs/integrations along with its potential/powerful use cases in generative AI (including but not limited to multimodality).

Bangalore Urban, Karnataka, India

I worked on large scale Machine Learning systems, serving 200 million+ users. My work involved building and integrating big data pipelines, training machine learning/deep learning models, and the associated prediction services. I am also proficient in deploying and migrating the above-mentioned services and pipelines.A brief intro to few of the projects I have deployed to production :- Built a live-stream ranking system from scratch. Used an ensemble approach involving graph based contrastive deep learning / user-user embedding similarity / user-content embedding similarity. This increased DAU by a factor of 20% and average watch percentage per viewer by a factor of 15%. - Created & deployed a classification model to identify potential high revenue generating users for Nostra , a gaming platform, this aided in targeting correct user cohorts. Increased conversion rate for ad-campaigns by a factor of 30%.- Built an AI tool to automatically identify and generate relevant short video snippets from recorded live stream videos. Utilised LLMs ( GPT-3.5) , faster-whisper, video-intelligence models for the same. Used this to capture user attention, this led to a 2x increase in engagement metrics. - Architected a real-time deep learning-based recommendation system for Roposo: a short video, and shopping app. Increased the engagement metrics by 3x.

Bangalore Urban, Karnataka, India

- Conducted Research on video models, completed my thesis on Genre Detection using Video Feature extraction.- Extracted video key frame segments from raw app video data and converted audio data to a spectrogram. Benchmarked various state-of-the-art Deep Learning models and achieved a higher accuracy by deconstructing /tweaking layers of a 3D-ResNext101 model. Resulted in +14% accuracy from the set baseline.- Created data visualization graphics, translating and aggregating complex big data into comprehensive visual representations. Curated a dashboard for the same which was used in showcasing company performance to key stakeholders and investors (For Ex - Google, Microsoft, Reliance etc).

India

I was in the founding team of hustlezy, an e-commerce website aiming to provide unique products from all around the world to indian customers.I played key roles in bootstraping the process, was involved in website UI/UX design, Google SEO, Facebook Marketing analytics and helping in building the marketing strategy and customer acquisition strategy. Generated revenue of 500,000 INR within the first 10 days. Current revenue is over 5,50,00,000 INR.Left the company to pursue interest in AI, still in touch with the founder, and sometimes advise on certain topics related to business growth.

New Jersey, United States

I helped analyse big data and build live dashboards for the company.Also deep dived into understanding explainable AI and worked on CAUSEv2 models that the company was building at the time.

Noida Area, India

I was given the task of understanding and working with core NLP concepts and using them to build a pipeline, I used keras and deep dived into neural netwroks and learnt how to build them as well.

Master'S Degree, Data Science, 4.0

Activities and Societies: Algorithmic trading : - working with distinguished traders/faculty to create AI based signals for momentum based.

Bachelor Of Technology, Computer Science And Communication Engineering, 8.51/10

Education record