Niyati Thole

Prompt Engineer at Equal.AI (2025-09 – Present)

Built and optimized production speech-to-speech LLM systems and designed multi-agent prompt architecture

• Built and optimized production speech-to-speech LLM system (Gemini, OpenAI) serving 200K+ DAU, achieving 95% response accuracy through 200+ prompt iterations and reducing critical failures in live call workflows
• Designed multi-agent prompt architecture (6+ participant agents) enabling language detection, intent parsing, tool execution, and contextual memory to improve conversational accuracy and flow at scale
• Built automated LLM evaluation infrastructure with Python-based pipelines, replacing manual QA and scaling call analysis 5x, doubling issue detection coverage
• Analyzed 1,000+ production call recordings to identify failure patterns (hallucinations, listening errors, language gaps), improving user retention by 30% and reducing call drop-offs through targeted system improvements
• Engineered and optimized prompts across Gemini, OpenAI, and Ultravox, tailoring for language support and model behavior while reducing token usage by 50% to improve inference cost and performance at scale

AI Researcher & Scriptwriter at Plivo (2025-03 – 2025-09)

Produced AI-focused scripts and content assets, designing end-to-end content strategies for AI use cases

• Produced 70+ AI-focused scripts and content assets, translating complex LLM concepts into user-centric narratives, driving 50%+ engagement and 30% subscriber growth
• Designed end-to-end content strategies for 20+ AI use cases, identifying user pain points and improving communication efficiency across internal and external stakeholders by 40%
• Conducted audience and sentiment analysis across platforms with a 25-member team, deriving insights that informed AI product positioning and improved campaign ROI

Content Designer at QuitSure (2024-07 – 2025-03)

Developed behavior-driven de-addiction program and led multi-channel user engagement strategy

• Developed a behavior-driven de-addiction program and redesigned website experience based on user insights, driving 40% higher engagement and 20% increase in traffic
• Led multi-channel user engagement strategy, boosting social media engagement by 35% through audience segmentation and data-driven content optimization

Bachelor of Psychology in Psychology – Pune University (2022 – 2024)

HSC Class XII – Pune University (2019 – 2021)