DeepHorus is an innovative tech media and intelligence platform focused on AI, blockchain, startups, and emerging technologies.
Published on

How Harsh Vardhan Goenka became founding engineer #1 at Kalpa Labs

Authors
  • avatar
    Name
    DeepHorus
    Twitter

Harsh Vardhan Goenka has joined Kalpa Labs, a Y Combinator F25 startup building generalist audio models, as founding engineer and researcher. He previously worked at Rubrik Inc., helping build Rubrik Cloud Vault from its early days. At Rubrik, he rewrote testing infrastructure so suites that took hours ran in minutes.

Goenka’s path came through founders Prashant and Gautam, whom he knew from undergraduate studies. The three co-founded a competitive programming club and often debated algorithmic and mathematical problems. Goenka initially declined a role but revisited the opportunity four months later when he wanted a change.

He interviewed around 11 p.m. on a spare Windows laptop and spent three hours solving a file-system problem. The experience gave him a rush he had not felt in months, he said. He described the outcome as “God’s plan.”

Kalpa Labs is pursuing audio’s missing “GPT-3 moment.” Instead of only text-to-speech, it wants one model that can voice a podcast from a script, sing in a user’s voice, or generate audio from detailed mood and tone prompts. Some capabilities are already unlocked.

Since joining, Goenka has moved from ML infrastructure toward ML research despite no prior expertise. He admitted feeling clueless at times, but said Kalpa Labs expects team members to stretch into unfamiliar territory. The company argues fresh domain expertise can bring perspectives specialists miss.

Goenka said early-stage engineering requires broad ownership, including design, product workflows, and user experience. He recalled late nights editing Kalpa Labs’ first launch video with a small team. They hid jokes in the final cut.

He compared startup life to a cosine curve: weeks of feeling stuck followed by breakthroughs and dopamine rushes. His advice is to recall wins and trust the team. He believes they will iterate and improve.