Infrastructure Engineer

SourcingXPress; Surat; On-site; INR 900,000 - 1,300,000; Full time

Job summary

VideoSDK is a developer-first real-time communication platform powering real-time video, audio, live streaming, and AI voice applications. We seek a Voice AI Infrastructure Engineer to build and run the infrastructure moving audio through speech recognition, an LLM, and speech synthesis in under a sub-second budget.

This is an infrastructure role, not a modeling role, requiring backend work with Go, Rust, or Python, and deployment on Kubernetes and Docker.

Qualifications

• 1+ years writing backend code that ran in production - internships and substantial side projects count.
• Solid in at least one of Go, Rust, or Python.
• Working knowledge of Docker; comfortable with Kubernetes or keen to get there fast.
• Some exposure to Redis and to observability tooling like Prometheus, Grafana, or OpenTelemetry.
• Genuine interest in real-time systems, where latency is a correctness problem rather than a nice-to-have.

Responsibilities

• Write services in Go, Rust, or Python that sit on the real-time audio path
• Deploy, scale, and debug on Kubernetes and Docker
• Use Redis for session state, pub/sub signaling, and rate limiting
• Instrument the pipeline with traces and metrics, including per-turn latency breakdowns (ASR -> LLM -> TTS)
• Chase down tail latency, jitter, and dropped sessions

Skills

Go

Rust

Python

Real-time systems

Education

BE/BTech in CS/IT

Tools

Docker

Kubernetes

Redis

Prometheus

Grafana

OpenTelemetry

Job description

Company: VideoSDK

Website: Visit Website

LinkedIn: Visit LinkedIn

Business Type: Startup

Company Type: Product

Business Model: B2B

Funding Stage: Seed

Industry: Information Technology

Job Description

Position: Voice AI Infrastructure Engineer

Location: Surat, Gujarat (Onsite)

Education: BE/BTech in CS/IT.

VideoSDK is a developer-first real-time communication platform building infrastructure for real-time video, audio, live streaming, and AI voice applications. We provide low-latency, scalable APIs and SDKs that power voice AI agents, conversational experiences, and real-time communication at scale, solving complex problems across real-time audio, WebRTC, media infrastructure, distributed systems, and cloud scalability.

The role

We build voice agents that talk to people in real time. A phone call has no loading spinner — every millisecond of latency is audible. You'll help build and run the infrastructure that moves audio through speech recognition, an LLM, and speech synthesis, and back again, inside a sub-second budget. This is an infrastructure role, not a modeling role.

What You'll Do

• Write services in Go, Rust, or Python that sit on the real-time audio path
• Deploy, scale, and debug on Kubernetes and Docker
• Use Redis for session state, pub/sub signaling, and rate limiting
• Instrument the pipeline with traces and metrics, including per-turn latency breakdowns (ASR -> LLM -> TTS)
• Chase down tail latency, jitter, and dropped sessions

What We're Looking For

• 1+ years writing backend code that ran in production - internships and substantial side projects count
• Solid in at least one of Go, Rust, or Python
• Working knowledge of Docker; comfortable with Kubernetes or keen to get there fast
• Some exposure to Redis and to observability tooling like Prometheus, Grafana, or OpenTelemetry
• Genuine interest in real-time systems, where latency is a correctness problem rather than a nice-to-have

Nice to have

• WebRTC, SIP/telephony, RTP, Opus
• LiveKit, Pion, or any streaming media work
• Anything you've shipped that had to stay up