Back to jobs

AI Engineer

Ken42Bengaluru, Karnataka, IndiaPosted 2w ago
AI Engineer
Apply on LinkedIn

Job Description

Who are we:

Ken42 is an automation-first operating system for higher education institutions. Through a single

unified platform, it digitally transforms and streamlines every aspect of institutional operations-from admissions and academics to fees, finance, and learning. Know more about us on https://ken42.com/

Ken42 is building an AI-first operating system for higher education — spanning admissions, learning, operations, finance, and institutional intelligence (KAI). We are already embedded within one of the largest, most complex education groups in the region — a multi-entity, multi-campus institution with significant national and international presence. Today, our footprint is strong but partial. The opportunity ahead is to expand this into a full-stack, multi-geography, multi-function transformation.


What we are looking for:

Role: AI Engineer

We are building a cutting-edge, fully self-hosted Voice and Chat AI Agent. Our core USP is an agent capable of advanced, multi-database reasoning (converting natural language to SQL, Neo4j/Cypher, and MongoDB aggregations) in both English and formal Hindi. We are moving away from expensive SaaS APIs to a fully in-house open-source stack.

We are looking for a highly adaptable, AI-native engineer with at least 2 years of hands-on experience who codes at lightspeed using modern AI agents. You will play a pivotal role in architecting, developing, and scaling our conversational agents, integrating them with telephony infrastructure, and delivering seamless human-like interactions.

Experience Required: 2-3 years of hands-on experience in AI engineering or applied data science.

Location: Bengaluru


What you’ll do:

  • Build Ultra-Low-Latency Voicebots: Integrate and optimize voice agents (working with providers and models like Ultravox) alongside self-hosted Speech-to-Text and Text-to-Speech models (like Kokoro or XTTS). Manage complex voicebot mechanics like real-time streaming, speech interruptions (barge-in), and turn-taking.
  • Telephony & SIP Integration: Connect our AI voice agents to telecommunication networks for seamless inbound and outbound calls using SIP servers (Plivo, Twilio, etc.).
  • Complex Agent Orchestration: Design, build, and scale advanced multi-agent workflows and multi-step reasoning pipelines using LangGraph or Google ADK.
  • NLP-to-Database Execution: Build and maintain secure data-fetching tools for the LLM to query SQL databases, Neo4j graphs, and MongoDB instances based on user intent.
  • Model Optimization: Help deploy, prompt-tune, and evaluate open-weight models (like Qwen3.6-35B-A3B and Ultravox v0.6 Llama 3.1 8B) tailored for conversation and data retrieval


Must Have Skills:

  • Experience: 2+ years of professional experience in AI/ML engineering, specifically building, deploying, and scaling LLMs, chatbots, or voicebots in production environments.
  • Telephony & Voice AI Infrastructure: Proven hands-on experience integrating voice AI with SIP servers (Plivo, Mcube). Familiarity with voice agent platforms/providers like Ultravox, and a deep understanding of Voice Activity Detection (VAD) and two-way audio streaming.
  • Agentic Orchestration: Strong practical experience orchestrating complex LLM workflows using LangGraph or Google ADK.
  • Spec-Driven Development & Claude Code/Codex: You don't just write code manually; you are a power user of Claude Code/Codex (or similar agentic IDE tools). You know how to write precise technical specifications (specs) and guide AI to generate, refactor, and test production-level code.
  • Strong Python Foundation: You are comfortable writing clean, asynchronous Python (FastAPI, asyncio) to handle real-time streaming data, concurrent voice/chat WebSockets, and SIP trunking integrations.


Bonus points if you have

  • Experience with real-time audio/voice frameworks (like Pipecat, WebRTC, or LiveKit) for low-latency streaming.
  • Tinkered with local LLM serving frameworks (vLLM, SGLang, Ollama, or Hugging Face Transformers).
  • Knowledge of Graph Databases (Cypher/Neo4j) or complex SQL/NoSQL query structures.
  • A GitHub portfolio showing projects where you orchestrated multiple AI tools to solve a complex conversational or database-reasoning problem.