Our Story

DraftTheRecord was founded by Adnan Sherif and Yash Ramchandani, an ML researcher and product manager who worked at Google on Gemini, one of Google's frontier AI efforts.

Their work sat at the edge of a new generation of AI: models that could understand text, audio, video, and images together, and produce better outputs by reasoning across all of those signals at once. At Google, Adnan and Yash worked with the latest techniques in model fine-tuning, reinforcement learning, and quantization, to make these systems not only more capable, but also fast and efficient enough to run in practical settings.

While building at that frontier, they saw a widening gap. AI techniques were advancing rapidly, but most products outside the frontier labs were not making use of the latest advancements: the kind that could unlock truly high quality output.

Before starting DraftTheRecord, Adnan and Yash also built infrastructure for the emerging generation of voice AI agents and were among the early builders of fully offline meeting note-taking tools for lawyers. After working across voice AI, legal workflows, and multimodal models, they decided to bring that experience directly into court reporting.

They left Google, joined Y Combinator, and began building a new kind of transcription platform for the legal record. DraftTheRecord combines frontier AI techniques with the real workflows of court reporters and transcriptionists: capturing proceedings, identifying speakers, formatting transcripts, generating rough drafts, and supporting the exact punctuation and layout preferences that legal professionals rely on.

The result is software built for the people who make the record. For transcriptionists, DraftTheRecord helps turn AI drafts into certified transcripts faster. For court reporters, it enables high-quality transcription even offline and produces high-quality AI rough drafts.