ML DA Intern

About the Role

This is a 6-month, fully in-office internship based at our Bangalore, India office.

Key Responsibilities

• Support the evaluation of ML models by assisting with structured listening tests, A/B comparisons, and basic error tracking.
• Help replicate model bugs under guidance, enabling researchers to debug and improve models.
• Assist in curating datasets by organizing and maintaining audio files from various sources (e.g., calls, synthetic, open source).
• Contribute to audio data annotation by labeling simple features like clipping, silence, gender mismatch, etc.
• Track and record observations during model testing cycles, highlighting potential regressions or changes.
• Collaborate with team members to maintain clean, well-organized QA documentation and evaluation logs.

Requirements & Tech Stack

• Passion for speech/audio technology and interest in machine learning quality workflows.
• Good listening and observation skills, with attention to detail.
• Comfortable using tools like Excel, Google Sheets, and working with structured data (e.g., CSV files).
• Basic familiarity with audio editing or listening tools (Audacity, Praat, etc.) or willingness to learn.
• Ability to write clear, simple feedback and document issues in structured formats.
• Good communication skills and the ability to collaborate within a fast-moving, cross-functional team.

What You'll Learn Being Part Of Sanas

• Hands-on experience with cutting-edge speech ML models in real-world applications.
• How to evaluate model quality across various dimensions like naturalness, intelligibility, and artifact detection.
• Techniques for error analysis and diagnosing issues from real customer calls and live model output.
• Best practices for audio data curation, annotation, and dataset hygiene.
• Exposure to A/B testing frameworks and structured listening test methodologies.
• How to collaborate with cross-functional teams including ML researchers, QA, engineering, and customer success.
• Insights into how machine learning quality impacts user experience and product performance.
• Opportunity to contribute directly to model improvement cycles and learn how feedback drives innovation.

What We Offer

• A mission-driven environment that values impact, ownership, and transparency.
• Competitive compensation and equity in a fast-growing startup.
• Opportunity to work on technology with real-world social value and global scale.
• Regular opportunities to present findings and influence product roadmap.
• Support for learning and development – courses, books, conferences, and mentorship.
• A collaborative, no-ego team culture focused on curiosity, speed, and learning.