Senior Software Engineer - Real Time Audio

Qualitate•New York, NY

1d•$160,000 - $270,000•Onsite

About The Position

Qualitate is building the AI-native continuous intelligence platform for enterprises, investment firms, and asset managers. We automate the entire primary research workflow, from study design and expert recruitment through AI-moderated interviews and insight synthesis, replacing weeks of manual research with high output signal in hours. Primary research remains slow and transactional for everyone involved. Investment and corporate strategy teams spend hundreds of hours coordinating intermediaries, scheduling calls, and synthesizing fragmented insights. Experts face repetitive outreach, screener calls, schedule wrangling, and one-off engagements that offer them little control or lasting value. Qualitate transforms this experience with an AI-native expert network platform that gives participants the flexibility to contribute on their own time while enabling research at unprecedented scale. Qualitate’s AI Moderator conducts thousands of conversations with experts each month, capturing forward-looking purchasing plans, competitive displacement, and real-time ROI signals. Our continuous intelligence platform then turns these conversations into quantified, time-series outputs tracking over 10,000 public and private companies, queryable instantly in natural language. Our customers include some of the world's largest enterprises, hedge funds, PE firms, credit investors, and venture capital firms. The Role Qualitate’s AI Moderator runs live, voice-based interviews with thousands of experts every month. The core of that experience is real time: audio streaming in, speech-to-text, and conversational turn-taking, all delivered reliably to users around the world. As a senior engineer focused on real-time and audio systems, you’ll own that live layer end to end. You’ll improve upon and ship the scalable, low-latency voice pipeline, including capture, streaming, speech recognition, model orchestration and synthesis. You’ll also iterate on the WebSocket infrastructure that keeps sessions stable at scale. You’ll own reliability, observability and graceful degradation for systems where a dropped packet is a dropped insight. This is a high-ownership role on a small team. You’ll work directly with the founding team to define technical strategy for our interview platform, not just execute tickets. We’re looking for someone with 5+ years of experience who has shipped real-time, user-facing products and thrives where latency, reliability and product quality all matter at once.

Requirements

5+ years of experience building software. Ideally some of it in a startup or fast-paced environment where you owned full features end-to-end.
Production real-time experience. You’ve shipped user-facing real-time systems built on WebSockets, WebRTC, streaming media or multiplayer, and you know how to keep them fast and available under load.
Audio and voice fluency. You’ve worked with audio pipelines, speech-to-text, text-to-speech or voice agents, and you understand latency budgets, buffering, codecs and turn-taking.
Strong backend skills. Node/TypeScript or similar; you design services and APIs for reliability, scale and observability.
Proficient with frontend technologies. Ideally React. You can build responsive, live interfaces, not just wire up forms.
Fluent with AI models and agentic coding tools. You use AI daily and understand how to build real-time experiences around LLMs. This is core to how we work.
High ownership, low ego. You work independently, communicate proactively, and are comfortable making architecture decisions and defending them in code review.

Nice To Haves

You’ve built voice AI or conversational agents in production, such as phone agents, meeting assistants or live transcription.
You’ve worked with speech to text providers such as Deepgram or AssemblyAI.
You’ve built desktop applications (Electron, Tauri or native).
You have a deep understanding of in-browser audio capabilities.

Responsibilities

Evolve the WebSocket infrastructure behind every live interview, from connection lifecycle, session state and message routing to backpressure and reconnection, so thousands of concurrent conversations stay stable.
Own the path from incoming audio to accurate, low-latency transcripts. Working with STT providers, browser audio APIs, and ffmpeg
Build systems that grow from thousands to tens of thousands of simultaneous sessions.
Set and hold the bar for uptime on systems where a dropped connection is a lost insight.