TixelJobs
S
Stark digital media services pvt. ltd.via muse

AI Engineer (Agentic & Generative AI)

MH, INPosted 1w ago
ML EngineerMid LevelFull-time

About the Role

You will design, build, and ship AI solutions for real client problems. Your work includes agents that run multi\-step business processes, LLM\-powered applications, voice and WhatsApp agents, and document\-intelligence pipelines. You will own features from prototype to production, working with our full\-stack and delivery teams.
PuneFull\-Time

**Responsibilities**
--------------------


* Build agentic AI systems: tool\-calling agents, multi\-step workflow automation, and human\-in\-the\-loop flows
* Develop GenAI applications using LLM APIs, RAG pipelines, vector databases, and prompt and evaluation frameworks
* Implement voice AI and WhatsApp/chat agents, including STT/TTS, telephony, and conversational flow design
* Integrate AI services with client systems through REST APIs, webhooks, CRMs, and databases
* Deploy and evaluate small or local language models where cost, privacy, or data\-residency requirements apply
* Build document digitization and OCR pipelines in collaboration with the team
* Monitor quality, latency, and cost in production, and iterate on hallucination control, guardrails, and accuracy
* Contribute to client demos, POCs, technical scoping, and documentation
* Use AI\-assisted development tools to ship faster and share good practices with the team
**Requirements**
----------------


* 2\+ years of hands\-on experience building AI/ML\-powered applications (not only coursework or notebooks)
* Strong Python; working knowledge of FastAPI or Flask
* Practical experience with LLM APIs (OpenAI, Anthropic, Gemini, or open\-source models)
* Built at least one RAG system or agent end to end, including embeddings and a vector store (pgvector, Pinecone, Qdrant, FAISS, etc.)
* Prompt engineering, structured outputs, and function/tool calling
* Git, REST APIs, JSON, SQL, and Docker basics
* Clear written and verbal communication; able to explain AI limits to non\-technical clients
Share