MulticoreWare - Full Stack Developer - Agentic Systems

Job Description :

We are looking for a Full Stack Developer: Agentic Systems to build the product layer for AI-native workflows.

This role focuses on turning LLMs, agents, memory, and external tools into reliable, production-grade user experiences. You will design and ship systems where agents can plan, execute multi-step tasks, recover from failures, maintain context, and deliver consistent value across sessions.

Responsibilities :

• Build end-to-end product features across frontend, backend, and AI integrations
• Design agent workflows that support planning, tool use, failure handling, and recovery
• Integrate LLMs, memory, RAG systems, and external tools into production systems
• Build real-time AI interactions using streaming, partial results, and low-latency responses
• Improve reliability, observability, fallback logic, and production behavior of AI workflows
• Collaborate with ML, backend, product, and design teams to ship features from concept to production
• Iterate on AI workflows based on user behavior, evaluation results, and observed failure modes
• Establish reusable patterns for building scalable agentic systems

Requirements:

• Strong full stack engineering experience across frontend and backend development
• Solid understanding of system design, APIs, and production-grade architecture
• Experience building with LLMs, RAG systems, agents, or AI-powered applications
• Ability to work through ambiguity and make pragmatic engineering decisions
• Strong ownership mindset with experience taking features from idea to production

Tech Stack & Skills:

Core Engineering:

• Next.js, Node.js, Python
• SQL and NoSQL databases
• API design and backend architecture
• Docker

AI & Agentic Systems:

• LLM integration using OpenAI, Anthropic, or open-source models
• Agent workflows, tool use, memory, or RAG-based systems
• Streaming responses and real-time AI interaction patterns
• Agent frameworks: LangChain, LlamaIndex, CrewAI, AutoGen, or similar
• Vector databases: Pinecone, Weaviate, Qdrant, Milvus, or pgvector

Production & Reliability:

• Observability, reliability, fallback handling, and debugging in production
• Experience with evaluation frameworks for LLM or agent performance
• Experience with workflow orchestration systems

Nice to Have:

• Familiarity with prompt engineering, retrieval strategies, and context management
• Experience building AI products beyond chat-based interfaces