Role & responsibilities :

- Develop and deploy full stack machine learning, deep learning, and generative AI solutions.

- Apply advanced methods such as ensemble models, representation learning, transfer learning, and retrieval-augmented generation (RAG).

- Design, build, and optimize LLM-based systems, including assistants, copilots, and automation workflows.

- Understand and apply AI/ML architecture patterns, model selection strategies, and technical standards in coordination with architect/platform team.

- Lead prompt engineering, fine-tuning strategies, embedding pipelines, and retrieval systems.

- Build components of agentic ai solutions like MCP servers, skills, agents.

- Integrate the AI solutions in web applications using nodejs, Reactjs and .Net.

- Collaborate on deployment and monitoring of solutions in cloud environments with performance, security, and cost considerations.

- Own end-to-end design and implementation of AI enabled cloud systems from problem definition to production deployment.

- Contribute to evaluation of trade-offs across model accuracy, latency, scalability, cost, and maintainability.

- Contribute to architect and maintain scalable training and inference pipelines on cloud platforms.

- Contribute to ensure AI enabled solutions meet enterprise-grade requirements for reliability, security, observability, and compliance.

Preferred candidate profile :

- 8+ years strong development experience with at least 3 years .Net web frameworks and React coding experience and about 3-4 years in AI/ML based solutions development.

- Technical background : Must have worked as Developer for web and cloud applications and possess strong knowledge of Software Development Framework.

- Experience and expertise in Generative AI & LLM Systems with good knowledge of LLM tools like OpenAI / Azure OpenAI / Anthropic / open-source LLMs, Lang Chain and Lang Graph, Vector databases : FAISS, Pinecone, Chroma. Expert in python with data engineering libraries like NumPy, Pandas.

- Good knowledge of typescript.