Loading organizations...
Podonos develops a specialized platform and tools designed to simplify and enhance the evaluation of artificial intelligence models, with a particular focus on Voice AI. Its solution provides streamlined, transparent, and reliable assessments, enabling AI teams to analyze models across critical parameters such as naturalness, pronunciation, emotion, and noise tolerance, often integrating a human-in-the-loop methodology for accuracy.
The company was founded in 2024 by Soohyun Bae, who serves as CEO. Bae, holding an ML PhD from Georgia Institute of Technology and having previously co-founded Bobidi as CTO, identified the pressing need for a robust and efficient evaluation layer in Voice AI development. This insight led to the creation of Podonos, aiming to address the complexities and inefficiencies in current model assessment processes.
Podonos serves AI teams and model builders seeking to refine their voice AI applications. The company’s vision centers on empowering these teams with comprehensive and transparent evaluation capabilities. Its long-term goal is to accelerate the improvement and ensure the reliability of Voice AI models, thereby shortening analysis and data iteration cycles for developers worldwide.
Podonos has raised $3.1M across 2 funding rounds.
Podonos has raised $3.1M in total across 2 funding rounds.
Podonos has raised $3.1M in total across 2 funding rounds.
Podonos's investors include Kevin Moore, KAIST Venture Investment Holdings, Naver D2SF, Agile VC, Capitalize VC, Expert Dojo, Serac Ventures, Sixty8 Capital.
Podonos has raised $3.1M across 2 funding rounds. Most recently, it raised $2.4M Pre-Seed in September 2025.
| Date | Round | Lead Investors | Other Investors | Status |
|---|---|---|---|---|
| Sep 9, 2025 | $2.4M Pre Seed | Kevin Moore, Serac Ventures | Kaist Venture Investment Holdings, NAVER D2SF | Announced |
| Mar 1, 2025 | $680K Seed | — | Agile VC, Capitalize VC, Expert Dojo, Serac Ventures, Sixty8 Capital | Announced |
Podonos is a North America-based startup specializing in voice AI model evaluation, offering a platform that delivers comprehensive performance assessments within 12 hours.[1][2][3] It combines a global network of over 150,000 human evaluators with proprietary AI automation to test aspects like pronunciation, emotion recognition, noise tolerance, and persona alignment, serving AI model developers (e.g., Resemble AI, Play AI, Sanas AI) and service providers across sectors such as healthcare, finance, gaming, and advertising.[1][2][3] The company recently raised $2.4M in pre-seed funding from NAVER D2SF, enabling rapid commercialization since its February 2024 founding and supporting multilingual evaluations in 9+ languages and 13 locales.[2][3]
Podonos solves the critical challenge of reliable, fast voice AI validation amid exploding AI model proliferation, where traditional methods like MTurk are slow and error-prone, often taking weeks.[3] Clients gain actionable insights for model improvement, benchmarking, and selection tailored to specific goals, environments, and audiences, accelerating development and deployment while meeting industry standards like ITU-R P.835.[2][3]
Podonos was founded in February 2024 by CEO Soohyun Bae, a serial entrepreneur and engineer with prior experience co-founding two ventures, including Ticktock.AI (sold to Niantic) and a Y Combinator-backed AI red-teaming startup focused on community-driven evaluations.[2] Bae's background in AI model expertise and workforce management directly shaped Podonos's human-in-the-loop (HITL) system, addressing gaps in quantitative speech naturalness assessment that remain rare globally.[2]
The idea emerged from the need for efficient "AI for AI" tools in the voice domain, with the team achieving product commercialization in just 18 months through engineering excellence.[1][2] Early traction came swiftly: the platform gained trust from leading voice AI startups like Play AI, which praised its 9-hour turnaround versus weeks of manual effort, and secured investment from NAVER D2SF to fuel expansion.[1][3]
Podonos rides the explosive growth of voice AI, fueled by generative models demanding rigorous, scalable evaluation to ensure quality in diverse real-world scenarios like noisy environments or personalized personas.[1][2] Timing is ideal amid AI proliferation, where developers and deployers need fast validation to iterate—Podonos fills a niche in "AI for AI" tools, as highlighted by NAVER D2SF's investments across the AI value chain.[1]
Market forces favoring Podonos include rising demand for hyperpersonalized, multilingual voice tech in sectors like healthcare (patient interaction), finance (secure auth), gaming (immersive NPCs), and advertising (targeted audio).[2] It influences the ecosystem by enabling trustworthy benchmarking, reducing deployment risks, and accelerating monetization, much like quality gates in software dev but tailored for audio AI.[3]
Podonos is poised for rapid scaling post-$2.4M pre-seed, with plans to expand into multimodal evaluations and deeper sector penetration, leveraging its HITL edge as voice AI integrates into everyday apps.[1][2] Trends like multimodal AI, global localization, and regulatory scrutiny on AI bias will amplify demand for its precise, fast insights, potentially positioning it as the go-to standard for voice model trust.
As AI models commoditize, Podonos's role in unlocking monetizable, high-performance voice tech will grow, empowering developers to deploy confidently worldwide—turning evaluation from bottleneck to superpower.[3]
