As a Data Science Intern at Zivy, you will be working on the core engine of classifying and prioritizing messages across different work tools to help our users cut through the communication noise. Your work will help people around the world discover the most important work in their daily jobs and create more impact for their company, their team and for themselves.

What will you be doing?

• Natural Language Processing: Develop, implement and use LLMs and machine learning models for processing millions of text messages in a day and classifying them into categories personalized to the user. Additionally, summarize and rank these test messages for the user.
• Parsing and extracting information out of documents and images for enabling good user experience on Zivy.
• Build new AI products and Agents for helping managers and leaders get their work done.
• Fine-tune and optimize existing LLMs and ML models, to enhance accuracy and performance. Build reinforced learning models which understand the user communication preferences and help them communicate better and faster.
• Conduct A/B testing and performance evaluation of ML models to drive continuous improvement
• Work with Backend Engineers and DevOps to Deploy, Maintain and Scale the ML models in production environment considering factors such as latency, cost etc.

What we’re looking for ideally

• Stay up-to-date with the latest advancements in machine learning and AI tools, frameworks, and libraries, and apply them to improve our ML capabilities.
• Experience with at least some of the following: Python, Scikit-learn, Pytorch, LLM APIs (like OpenAI/Claude), Cloud Provider like AWS, Database (SQL/NoSQL), Git, Basics of DSA, System Design, AI Agents.

Hiring process

1. If shortlisted, one of our engineers will have an exploratory call with you where we’ll talk more about Zivy, the team, the role, etc. Post that, we’d love to get to know more about you, your past work experience and a bit of chit-chat.
2. Post the exploratory call, if we mutually agree that we could be a good fit, we’ll have around an hour of technical round where we’ll discuss relevant technical topics focused on evaluating your depth of knowledge in machine learning theoretical concepts, including but not limited to natural language processing (using both traditional NLP and LLMs), model fine-tuning, data preprocessing, and building products using AI agents.
3. The third round will be a coding challenge to assess your expertise with the machine learning tech stack. This round will test your ability to write clean, efficient, and well-structured code for machine learning tasks, including model implementation and optimization.

We may combine the 2nd and 3rd round into a single round. 4. If things go well, one of the founders will have a final round — this is just for a vibe check, culture fit :)

We will ensure the entire process is completed within 2 weeks, generally within a week. In case things don’t work out, please be sure it’s not a reflection of you but just not the right match at this point of time. It’s always frustrating to not receive any feedback / why not the right match / why not the right time. We will always provide detailed feedback!

Useful Links

Why are we building Zivy?