Software Engineer - 2 (Voicebot)
Job Fit Check
Base Career helps you apply smarter for this job.
Key skills for this role
Role Overview
The Voicebot team builds and operates a real-time voice AI product handling live phone conversations for enterprise customers.
The role covers the model layer and live conversation experience, including fine-tuning, evaluations, speech quality, latency, and turn-taking.
The engineer will build, evaluate, and ship changes that improve call quality and business metrics in live deployments.
Key Skills for This Role
Full Job Posting
About Exotel
Exotel provides AI-powered communication solutions for enterprise customer engagement across voice, agents, bots, and omnichannel interactions.
The company supports live customer conversations for enterprise clients across industries including BFSI, logistics, healthcare, education, and e-commerce.
Voicebot Team and Role
The Voicebot team builds and operates a real-time voice AI product handling live phone conversations for enterprise customers.
The role covers the model layer and live conversation experience, including fine-tuning, evaluations, speech quality, latency, and turn-taking.
The engineer will build, evaluate, and ship changes that improve call quality and business metrics in live deployments.
Responsibilities
- Build and maintain LLM and speech evaluation frameworks for model and prompt changes.
- Run fine-tuning experiments using full fine-tuning, PEFT, LoRA, and QLoRA on open-weight models.
- Benchmark LLM and ASR/TTS engines across providers and self-hosted options using cost, latency, and quality data.
- Diagnose production conversation failures such as poor turn-taking, misrecognition, latency spikes, and prompt regressions.
- Ship changes into the live conversational pipeline with instrumentation and alerting.
- Own design, evaluation, deployment, and monitoring across the software development lifecycle.
Must-Have Requirements
- Strong grounding in artificial neural networks and transformer architecture, including attention, tokenization, and decoding strategies.
- Hands-on experience with LLM evaluation harnesses, LLM-as-judge setups, or regression suites.
- Hands-on fine-tuning experience using PEFT, LoRA, or QLoRA on an open-weight model for a real task.
- Working knowledge of speech and ASR/TTS evaluation, including WER, latency, and diarization.
- Strong Python skills for production engineering.
- Two to four years of software or ML engineering experience with production-system experience.
- A track record of shipping and owning changes in production.
- Strong analytical rigor and a measurement-driven approach.
Good-to-Have Qualifications
- Experience with real-time audio or streaming systems and streaming-versus-batch tradeoffs.
- Experience with agentic orchestration and LLM tool-calling patterns.
- Exposure to retrieval-augmented generation, embeddings, vector stores, and retrieval strategies.
- Familiarity with self-hosting or serving open-weight models.
- Familiarity with observability for AI workloads, including cost tracking and quality dashboards.
- Experience with multi-tenant SaaS constraints.
- Prior experience in voice AI, IVR, or contact-center domains.
Ways of Working
- The team expects engineers to build, evaluate, ship, monitor, and remain accountable for live systems.
- Model and prompt changes are expected to have an evaluation story before release.
- The role involves collaboration with voicebot architecture, product, and field delivery engineers.
Apply for this job in 1 click
Skip the repetitive application forms
Install the Base Career Chrome Extension and autofill job applications across major job boards with your profile.
Trusted by over 500,000 job seekers on Base Career