Role Overview :

We are looking for an Applied Scientist (Agentic AI) with hands-on experience in building and deploying production-grade chat, voice, and multimodal AI agents.

In this role, you will design autonomous agent workflows, state/memory management architectures, LLM integrations, and real-time voice pipelines tailored for high-impact Real Estate and PropTech automation systems.

Key Responsibilities :

1. Autonomous Agent Design & Architecture :

- Build multi-step, autonomous agent architectures using stateful memory, structured tool calling, and agentic loops.

- Implement task execution systems where agents dynamically plan tasks, invoke APIs/MCPs, and verify outcomes.

- Manage persistent context windows, session state, and vector databases (RAG) across long-running interactions.

2. Multimodal & Voice Pipeline Engineering :

- Architect end-to-end, low-latency conversational voice pipelines including ASR, TTS, and active interruption handling.

- Integrate voice, text, image, and structured metadata processing for complex property discovery and lead qualification.

- Fine-tune models for Indian languages, multi-accent handling, and code-switching (e.g., Hinglish).

3. MarTech & Sales Automation :

- Develop persona-driven agents for dynamic content generation and automated follow-up workflows.

- Automate intelligent lead scoring, scheduling, and real-time data sync into CRM platforms.

4. Observability & Security :

- Design robust moderation systems and guardrails to minimize hallucinations and toxicity.

- Implement comprehensive logging, tracing, and telemetry frameworks to monitor AI behavior in production.

5. Technical Execution :

- Transition experimental agent architectures into production-ready, scalable microservices.

- Partner with Product, UX, and Backend Engineering teams to iterate on latency and quality.

Required Qualifications :

- Education : B.Tech / Master's / PhD in CS, AI, ML, or Data Science (min. 60% academic record).

- Experience : 4 - 11 years in AI/ML with expertise in LLM applications, conversational agents, and RAG.

Tech Stack :

- Frameworks : LangChain, LlamaIndex, AutoGen, CrewAI.

- LLMs & Infrastructure : OpenAI API, Anthropic, open-source models (Llama, Mistral), Vector DBs (Pinecone, Weaviate).

- Voice : Twilio, Exotel, WebRTC, Whisper, ElevenLabs, Deepgram.

- Backend : Python, FastAPI, Docker, AWS/GCP/Azure.