Ai Experience Architect

**Domain: AI Interfaces | LLM Behavior Design | Multimodal UX | Voice Systems**

**About the Role**

We’re seeking a visionary **AI Experience Architect** to design how humans engage with intelligent systems — across voice, screen, camera, and physical environments.

In this role, you won’t just write prompts or connect APIs — you’ll **craft the behavior, tone, logic, and flow** of advanced AI agents that **guide, teach, respond, and adapt** to users across complex real-world journeys.

Think **LLM-driven personalities**, **voice-first interfaces**, **emotion-aware AI**, and **real-time feedback systems**. Your work will power conversational flows and intelligent behaviors that feel **natural, empathic, and deeply human**, yet entirely autonomous.

**What You’ll Do**
- **Architect AI-driven interaction journeys** across voice, text, and screen.
- Design **modular prompt chains**, memory structures, and LLM workflows for different use cases (e.g., guide, tutor, agent, counselor).
- Build and test **AI personas** with distinct tones, teaching styles, and escalation logic.
- Create interaction maps and branching logic for real-world use cases (not just chatbots).
- Integrate multimodal inputs (voice, face, behavior) into response flows.
- Collaborate with CV, voice AI, UX, and backend teams to synchronize behavior.
- Continuously **test and refine user experience** across devices and scenarios.
- Implement guardrails, context windows, and fallback logic for edge cases.

**Key Skills & Experience**

You don’t need to tick every box — but we’re looking for those who can combine LLM know-how with UX vision and behavioral design.

**Required**:

- 2-5 years experience designing with **LLMs (GPT, Claude, Mistral, etc.)**:

- Proven skill in **prompt engineering**, agent design, and tool/plugin orchestration
- Deep understanding of **context management, instruction tuning**, and LLM limitations
- Strong grasp of **voice-AI workflows** (Whisper, TTS, STT, latency tuning)
- Experience crafting **AI personalities** and conversational tone systems
- Ability to design **flow diagrams, user journeys**, and real-world escalation logic

**Bonus**:

- Built intelligent agents that interact with people via **voice or physical interfaces**:

- Experience with **emotion detection**, behavior tracking, or adaptive learning
- Background in **game narrative design**, **interactive storytelling**, or **conversational UX**:

- Experience working with tools like LangChain, RAG pipelines, vector stores

**You Might Be a Fit If You**:

- Think of AI as a **living teammate**, not a tool.
- Love prototyping real-world behavior (not just writing responses).
- Understand how **context, memory, and tone** shape trust in AI systems.
- Can bridge the gap between **technical LLM pipelines** and **empathic user experience**.
- Have worked on chatbots, agents, games, IVR, or AI tutors — and wanted to go **much deeper**.

**Perks & Culture**
- **Build a system that talks, teaches, and learns — for real**:

- Work with **LLM, CV, IoT, and edge-AI engineers**:

- You’ll have **creative control** over AI behavior at scale
- Shape the **future of human-AI interaction in physical space**:

- Think like a product owner, not just a task executor

**Before You Apply**

We’re not hiring “prompt writers.”

We’re looking for **architects of intelligence** — people who think in flows, feedback loops, and frictionless user emotion.

**Application Instructions**
- Send us your **portfolio or demo** — voice bots, LLM projects, or any interaction systems
- Include a short note on:_“If you could design a voice AI for a physical space, what would it sound like? What would make it magical?”_

Pay: ₹13,000.00 - ₹44,000.00 per month

Ability to commute/relocate:

- Sardar Nagar, Rajkot, Gujarat: Reliably commute or willing to relocate with an employer-provided relocation package (preferred)

Application Question(s):

- Describe one intelligent agent or chatbot you’ve built.
- What model did you use?
- How did you design its behavior/personality?
- How did it handle memory, escalation, or fallback?
- Share a real prompt (or chain of prompts) you wrote for a complex task using an LLM.

Briefly explain why it worked well, and how you refined it
- What are the 3 most important ingredients in crafting a believable AI personality?
- Imagine designing an AI receptionist for a public space. It has a camera, speaker, and large screen.

How should it behave when:

- A child approaches?
- A confused adult approaches?
- No one is nearby?
- A user keeps repeating the same phrase but the AI doesn’t understand.

How should your system respond — both in message and logic?
- Which of the following tools or frameworks have you used in real-world projects?

(Select all that apply)
- GPT / Claude API
- LangChain or Semantic Kernel
- Whisper / STT / TTS libraries
- Pinecone / Weaviate / FAISS
- Dialogflow / Rasa / Voiceflow
- Emotion detection (DeepFace, MediaPipe, etc.)
- Figma / Framer