Aug 2025 – Present · 10 mos

• Integrated GPT-powered title/description/script generation and automated background-music selection (S3), paired with a Slack review workflow and staged rollout — reduced average creator time-to-publish by 40% and boosted adoption of suggested metadata by 35%.
• Proposed and designed Text-to-Speech (TTS) and Speech-to-Text (STT) features with complete architecture and documentation, enhancing accessibility and user engagement by 15%.
• Established a background processing pipeline (OpenAI Whisper → GPT) to transcribe podcasts and
• auto-generate highlight clips — cut highlight generation time by 80% and drove highlight plays up by 50%.

May 2025 – Aug 2025 · 3 mos

• Backend Development:-
• Developed and maintained scalable backend microservices using Golang and TypeScript, improving response times by 30% and optimising core modules like search and comments, and playlist.
• Developed a Node.js script to optimise user feed by automatically removing outdated and unwanted fields from Amazon Neptune DB entries older than 90 days, and scheduled the task using a cron job on AWS EC2, enhancing data relevance and system performance.
• Constructed a basic version (PoC) of an AI assistant using Langflow and AstraDB to test the idea of a “best friend” chat experience, aiming to improve user interaction and engagement.
• Added Google Image Search support, so users can find and add related images when creating a voicepod or voicepool, making their content more engaging.
• Tech stack: Golang, Typescript, GraphQl, DynamoDB, NeptuneDB, S3, Redis.