Job Description & How to Apply Below
Location:
Chennai (Hybrid)
Company:
Allsec AI
Website:(Use the "Apply for this Job" box below)..ai
Engagement Type:
Full-Time | Early-Stage Start-up
AI/ML Engineer Voice AI
• Conversational Systems
• Execution & Optimisation
About Allsec AI
Allsec AI is an AI-native, digital-first, omni-channel CX managed services company. We help enterprises unlock scalable, efficient, and real-time customer experience operations through agentic AI, intelligent workflows, and automation.
We are redefining outsourcing by delivering packaged, outcome-led services, with full ownership of delivery outcomes, not just technology.
About This Role
The core engineering challenge of this role is to create AI voice-based CX within a trifecta of cost, quality, and latency. You will execute on decisions that balance all three (across voice models, language models, and real-time pipelines) in a production CX environment where those trade-offs have direct business consequences.
You will be working on a live, multi-model AI platform where your responsibility lies in contributing meaningfully to making it better, faster, and more economical, under the technical direction of the Principal Engineer. You will be expected to ship, optimise, debug, and iterate, often across layers of the stack you may not have originally written.
Beyond the core platform, you will support the Principal Engineer and operational leadership team in developing new use cases, building out conversation flows, and contributing to the product’s expansion into new verticals.
If you are someone who thinks about AI not just as a technical system but as something that has to work for a real person on the other end of a phone call, this role is built for you.
Key Responsibilities Voice AI & Model Optimisation
• Work on and improve the end-to-end voice pipeline — STT, TTS, interruption handling, silence detection, and turn-taking logic.
• Execute on optimisations across the cost-quality-latency trifecta: implement caching, streaming, and model selection decisions under technical guidance.
• Improve STT accuracy across accents, call environments, and language variations through systematic testing and tuning.
• Evaluate new voice and language models and provide recommendations to the Principal Engineer on provider selection.
LLM Integration & Pipeline Engineering
• Build, maintain, and improve LLM pipelines including RAG, prompt engineering, and context management.
• Work across multiple LLM providers — benchmarking based on cost, latency, and output quality as directed.
• Contribute to evaluation frameworks that measure model performance across cost, quality, and latency dimensions.
• Manage and improve vector databases, embedding pipelines, and retrieval systems that power the knowledge layer.
Use Case & Conversation Support
• Support the translation of client requirements into conversation flows and call scripts, working alongside the operational and business team.
• Build and refine branching logic for multi-turn voice conversations — handling edge cases, fallbacks, and escalation paths.
• Contribute ideas and execution on new use cases as the business grows.
• Assist the Principal Engineer in extending existing flows into new industry verticals — from BFSI to healthcare, retail, and beyond.
Platform & Infrastructure
• Be operationally self-sufficient — deploy, monitor, and debug your own work across the existing multi-cloud stack (Digital Ocean, Azure, GCP).
• Contribute to basic UI/UX improvements where the AI layer meets the client-facing product.
• Participate in code reviews, maintain clean and documented code, and work within engineering standards set by the Principal Engineer.
Requirements
• 4+ years of software development experience — strong development fundamentals are non-negotiable; you must be able to write clean, production-quality code.
• Hands-on experience with LLM APIs and integration (OpenAI, Anthropic, Gemini, or equivalent) in a live product environment — not just experimentation.
• Familiarity with RAG pipelines, vector databases (Pinecone, FAISS, or similar), and prompt engineering.
• Enough cloud and Dev Ops competence to own your own deployments — you do not need…
Note that applications are not being accepted from your jurisdiction for this job currently via this jobsite. Candidate preferences are the decision of the Employer or Recruiting Agent, and are controlled by them alone.
To Search, View & Apply for jobs on this site that accept applications from your location or country, tap here to make a Search:
To Search, View & Apply for jobs on this site that accept applications from your location or country, tap here to make a Search:
Search for further Jobs Here:
×