Infer Hires Engineers in US and India to Scale Voice AI Platform

Infer, a Y Combinator S21 startup based in San Mateo, California, builds personalized AI voice agents designed to handle lead qualification for insurance and financial services. The company is currently expanding its technical team with open engineering roles in both California and Bengaluru, India.

The startup's voice agents replace standard web intake forms with conversational audio interactions. Rather than presenting prospective buyers with static fields, the agents answer complex domain questions around insurance terms such as deductibles and copays before completing a warm transfer to human insurance agents. Speaking on the AI Minds podcast hosted by Deepgram, co-founder Vaibhav Saxena explained that roughly 80% of customer interactions in inbound insurance calls consist of repetitive queries. Infer uses voice agents to automate these interactions without dropping conversion rates.

From Audio Summarization to Voice Automation

Infer was founded by Vaibhav Saxena, Suneel Matham, and Urvin Soneta. Saxena transitioned into deep learning research at Purdue University after training as an architect. Matham holds a BTech in Mechanical Engineering from the Indian Institute of Technology Madras and completed the Plaksha AI fellowship, while Soneta studied economics at Ashoka University.

The trio originally entered Y Combinator's Summer 2021 cohort building Synth, an app focused on capturing, summarizing, and indexing audio from meetings, podcasts, and video. After observing low revenue growth and distribution challenges with horizontal audio tools, the team spent weeks analyzing recorded call logs from prospective enterprise customers. The prevalence of repetitive call scripts prompted the founders to pivot directly into verticalized AI voice agents for insurance lead qualification.

Personalization and Latency Benchmarks

During a follow-up interview on Deepgram, Saxena discussed how Infer runs controlled experimentation across speech traits to identify what drives customer engagement. The company tests variables including voice pitch, accent, prosody, and conversational speed while keeping call scripts constant. Their internal evaluations track metrics such as call duration and conversion rates across different age brackets and locations.

To optimize these interactions, Infer requires real-time streaming architectures that minimize latency across speech-to-text (STT), large language model (LLM) reasoning, and text-to-speech (TTS) pipelines.

Engineering Roles Across San Mateo and Bengaluru

Infer is recruiting across multiple technical domains to support customer deployments and platform scalability:

• In the US, Infer is hiring a Forward Deployed Engineer ($110,000 - $180,000) in San Francisco, CA. The role involves managing client integrations end-to-end, gathering workflow knowledge from insurance agencies, and converting operational practices into production system architecture.
• In Bengaluru, India, the team is hiring an Applied AI Engineer (₹2,000,000 - ₹5,000,000) to build evaluation infrastructure that benchmarks speech transcription, LLM reasoning, and tool utilization while running automated red-teaming tests.
• Also in Bengaluru, Infer is filling an Engineering Lead role (₹4,000,000 - ₹6,000,000) to direct technical strategy, mentor engineers, and set streaming reliability standards across backend platforms.
• A Senior backend Engineer position (₹2,000,000 - ₹5,000,000) in Bengaluru focuses on reducing pipeline latency across STT, LLM, and TTS systems while building observability tools for concurrent voice sessions.

See all open roles at Infer