Headline

Builder @Murf AI . Ex-Sharechat, OYO I IIT Kanpur

About

AI Product Manager with building Voice AI, Text-to-Speech APIs, Voice Agents, AI Dubbing solutions, and large-scale recommendation systems. Passionate about transforming cutting-edge AI research into products that create meaningful customer and business impact.

Location

Bengaluru

AI stack

anthropic apis

claude code

claude

cursor

elevenlabs

grok

Work experience

Product Manager at Murf AI, Oct 2024 - Present · 2 yrs, Bangalore Urban, Karnataka, India, Building AI-powered voice products for developers and enterprises, with a focus on Text-to-Speech APIs, Voice Agents, AI Dubbing, and Voice AI. 1. Own the product roadmap for Murf's AI platform, collaborating with engineering, AI researchers, design, and GTM teams to deliver high-impact features. 2. Lead the development of Text-to-Speech APIs, improving the developer experience through intuitive APIs, SDKs, and documentation. 3. Drive product strategy for Voice Agents and AI Dubbing, enabling businesses to build multilingual and conversational AI applications. 4. Worked closely with enterprise customers to identify product gaps, influence roadmap priorities, and deliver scalable Voice AI solutions. 5. Define product requirements, success metrics, and experimentation strategies while balancing customer needs with business objectives. 6. Architected and launched Murf’s multi-stage LLM Translation Agent for AI Dubbing featuring context-aware translation, a Judge layer for meaning/coherence checks, and a duration adaptation loop , improving translation accuracy from 30% to 92% (MQM score); published methodology and benchmark findings: murf.ai/blog/ai-dubbing-translation-quality-benchmark Blog Page : https://murf.ai/blog-author/ashutosh-singhania

Product Manager at Eloelo, Jan 2024 - Oct 2024 · 10 mos, Bangalore Urban, Karnataka, India, 1. Owned product roadmap for the LiveFeed Recommendation team and led development of a new live recommendation feed algorithm, resulting in a 10% relative retention gain. 2. Implemented new features including bounce rate, meaningful watch, slot level mechanism, and weights optimization of feed score to enhance user experience and drive product adoption. 3. Established an A/B testing framework to validate product hypotheses; oversaw feed algorithm for Async (Status) product using a mix of popularity and topical content signal

Associate Product Manager at ShareChat, Mar 2023 - Jan 2024 · 11 mos, Bengaluru, Karnataka, India, 1. Drove adoption of MMoE-based recommendation model and launched a two-tower candidate generation model,improving user retention by 5% and significantly boosting retrieval efficiency. 2. Initiated real-time training with a Monolith model in the recommendation feed and developed a novel weight assignment technique for multi-headed models, contributing to an additional 1.2% retention improvement. 3. Directed extensive multi-variant A/B testing and managed multi-million dollar budgets; launched new categories like Memes and Gaming driving DAU gains.

Senior Product Analyst at ShareChat, Jul 2021 - Mar 2023 · 1 yr 9 mos, Bengaluru, Karnataka, India, Moj Feed AI

Delegate at Harvard Project for Asian and International Relations (HPAIR), Jan 2021 - Jan 2021 · 1 mo

Data Scientist at OYO, Sep 2019 - Jul 2021 · 1 yr 11 mos, Gurugram, Haryana, India, 1.Handling Dynamic Pricing Model of South East Asia and Middle East countries 2.Designed a streamlined version of dynamic pricing codes which reduces run time from 3 hours to 6 mins. 3.Developed robust data driven matching models (Using T-test, correlation cofficient, matching market,P-value and other statistical metrics) and causal impact methodology to see the revenue impact using A/B testing and also won Employer of the month award. 4. Created inhouse Experimentation tool using R shiny Application to reduce the manual work. 5. Worked on Machine Learning(XGBoost,Random Forest,KNN,SVM) and deep learning models to find out important features which affects the occupancy and prices of hotels. 6. Built Competition Pricing Framework for China and South East Asian Countries using Ball tree Algorithm (Unsupervised Algorithm) to find the best competitor using amenities and price with the accuracy of ~95%. 7. Used Glove embedding score (NLP) to have the text analytics of hotels descriptions and find the similarity score. 8. Worked on Recommendation Engine which recommend prices for new properties .

Education

Minor, Industrial and Management Engineering, INDIAN INSTITUTE OF TECHNOLOGY KANPUR

Bachelor of Technology - BTech, Materials Science, INDIAN INSTITUTE OF TECHNOLOGY KANPUR