Aditi Ravishankar
Conversational AI @ Nurix (via Recro) | PGCert Data Science - IISc | Ex-Accenture
Bengaluru
3+ years of work experience
Oct 20253 mosDec 2025
Bengaluru, Karnataka, India
• Owned and designed human-like conversational user experience and prompt architectures for voice and chat AI agents at Nurix AI, for EdTech and Insurance clients.
• Led client-facing discussions to translate requirements into natural conversation flows, personas, guardrails, tool usage, and post-call analysis logic. Drove iterative refinements to align agent behavior with real-world user expectations.
• Independently delivered and optimized live AI agents by curating knowledge bases, cleaning noisy lead and call data, and tuning agent behavior. Reduced and optimized prompts, resulting in a 60% reduction in per-conversation-turn cost, while maintaining conversational quality.
• Improved post-call transcript analysis accuracy by ~50% through prompt refinement and built SQL-based analytics dashboards in Metabase to help stakeholders monitor call outcomes and agent performance.
• Actively tested and validated AI agents through call simulations, transcript reviews, and edge-case analysis, using feedback to improve conversational quality, accuracy, and reliability before and after go-live.
• Worked in a fast-paced, dynamic startup environment with high ownership and minimal supervision.
Feb 20241 yr 9 mosOct 2025
San Francisco, California, United States
• Contributed to multiple projects utilizing Reinforcement Learning with Human Feedback to enhance the performance of Large Language Models, including Google’s Gemini and Meta’s Llama.
• Worked on re-writing and refining code responses generated by the models in Python, Java, and SQL, significantly improving code efficiency and performance.
• Engaged in prompt engineering by designing complex prompts to challenge the model and induce deviations in code responses, which were then identified and corrected. This process was instrumental in generating high-quality training data, essential for improving model accuracy and performance in software engineering and data science.
Sep 20211 y 9 moJun 2023
Bengaluru, Karnataka, India
• Worked on developing a campaign management tool for Google marketers that simplifies workflow processes such as contract creation, approval, media requests, etc. pertaining to their projects.
• Worked as part of a team of 6 as a Full Stack Engineer, involved in both backend and frontend code development of the application in Agile Development methodology.
• Primary responsibilities included code development, customer requirement analysis, unit and integration testing, and maintenance.
• Further responsibilities included customer interaction, weekly manual production releases, documentation, etc.
• Used to lead some feature developments - Was responsible for managing, driving analysis, setting the technical direction, and coordinating with some team members for the successful delivery of some features of the product.
Apr 20203 mosJun 2020
Remote
Extracted transcripts from YouTube’s tech videos, cleaned and labeled the data to create a dataset, and used NLP and Machine Learning to obtain a model for classifying the videos into four categories (Reviews, Unboxing, Comparisons, and Others) with a maximum accuracy of 96%
Aditi’s AI Stack
Education
Postgraduate Certification, Computational Data Science
Indian Institute of Science (IISc)
20241 yr2025
Bachelor of Engineering - BE, Electronics and Communication Engineering
RNS Institute of Technology - India
20174 yrs2021
Join Aditi and 10k+ other professionals on
Tal is the app where bosses meet engineers & builders directly, without middlemen
Download Tal Boss