J<\/span><\/span><\/b>
<\/p>Job Description:\nJava Developer \u2013 AI Integration (Mid\-Level)<\/span><\/span><\/u><\/b>
<\/p>Role Overview<\/span><\/span><\/u><\/b>
<\/p>We are looking for a\nMid\-Level Java Developer (4\-9 years) with hands\-on experience in AI integration\nto join our product engineering team. In this role, you will be responsible for\nembedding AI capabilities \- including converting the user\-provided natural\nlanguage inputs, or database metadata information into valid product\nconfigurations, Retrieval\-Augmented Generation (RAG), LLM\-powered features, and\nagentic workflows \- directly into our Java/Spring Boot\-based enterprise\nproduct. You will work at the intersection of traditional backend engineering\nand modern AI/LLM technologies, owning the design, development, and maintenance\nof AI\-powered components end\-to\-end.<\/span><\/span>
<\/p>Key\nResponsibilities:<\/span><\/span><\/u><\/b>
<\/p>AI Feature\nDevelopment<\/span><\/span><\/u><\/b>
<\/p>- Capture user intent via a conversational/chat\n interface and translate it into valid product configurations.<\/span><\/span>
<\/li>- Extract metadata from heterogeneous sources \-\n file formats (CSV, Excel, PDF), relational database schemas, and API\n contracts (OpenAPI, WSDL) and convert them to valid product\n configurations.<\/span><\/span>
<\/li>- Design and implement RAG pipelines that retrieve\n contextually relevant information from internal knowledge bases and\n surface it through LLM\-generated responses.<\/span><\/span>
<\/li>- Integrate LLM\-powered features such as\n intelligent chat, document summarization, content generation, and semantic\n search into the existing product.<\/span><\/span>
<\/li>- Build and maintain AI agent workflows \- including\n tool use, multi\-step reasoning chains, and orchestrated LLM calls \- to\n automate complex product\-level tasks.<\/span><\/span>
<\/li><\/ul>- Author, version, and iterate on system prompts\n and user prompt templates to reliably steer LLM behavior.<\/span><\/span>
<\/li>- Apply advanced prompting techniques:\n chain\-of\-thought, few\-shot examples, output format enforcement, and\n instruction following.<\/span><\/span>
<\/li><\/ul>- Integrate with multiple LLM providers \-\n OpenAI/Azure OpenAI, Anthropic Claude, AWS Bedrock, and open\-source models\n (LLaMA, Mistral, etc.) \- through their APIs.<\/span><\/span>
<\/li>- Build provider\-agnostic abstraction layers using\n Spring AI and LangChain4j to allow model switching without core code\n changes.<\/span><\/span>
<\/li><\/ul>Backend\nEngineering<\/span><\/span><\/u><\/b>
<\/p>- Write clean, maintainable, production\-grade Java\n code following SOLID principles and enterprise design patterns.<\/span><\/span>
<\/li>- Integrate AI services with existing relational\n databases via JPA/Hibernate, including schema design for storing\n conversation history, embedding metadata, and audit trails.<\/span><\/span>
<\/li>- Manage build pipelines and dependency governance\n using Maven or Gradle.<\/span><\/span>
<\/li><\/ul> <\/span><\/span>
<\/p>Testing &\nQuality<\/span><\/span><\/u><\/b>
<\/p>- Write comprehensive unit, integration, and\n contract tests for AI\-integrated components.<\/span><\/span>
<\/li>- Handle non\-deterministic LLM output gracefully \-\n implement fallback logic, retry strategies, and output validation.<\/span><\/span>
<\/li>- Monitor AI feature performance in production;\n instrument logging and tracing for LLM calls, including token usage,\n latency, and error rates.<\/span><\/span>
<\/li><\/ul> <\/span><\/span>
<\/p>Mandatory Skills\n& Experience<\/span><\/span><\/b>
<\/p>Java & Backend<\/span><\/span><\/u><\/b>
<\/p>- 4\u20139 years of professional Java development\n experience.<\/span><\/span>
<\/li>- Strong proficiency in Spring Boot<\/b> \- REST\n controllers, service layer design, dependency injection, and configuration\n management.<\/span><\/span>
<\/li>- Experience building and consuming RESTful\n Microservices<\/b>.<\/span><\/span>
<\/li>- Working knowledge of JPA/Hibernate<\/b> \-\n entity modelling, JPQL, transaction management.<\/span><\/span>
<\/li><\/ul> <\/span><\/span>
<\/p>AI / LLM\nEngineering<\/span><\/span><\/u><\/b>
<\/p>- Hands\-on experience integrating with at least one\n major LLM provider API (OpenAI, Azure OpenAI, Anthropic, or AWS Bedrock)\n from Java.<\/span><\/span>
<\/li>- Practical experience with Prompt Engineering<\/b> \- not just calling APIs, but crafting, testing, and iterating prompts to\n achieve reliable, structured outputs.<\/span><\/span>
<\/li>- Experience building RAG pipelines<\/b> \-\n document ingestion, chunking strategies, embedding generation, similarity\n retrieval, and context injection into prompts.<\/span><\/span>
<\/li>- Working knowledge of LangChain4j<\/b> and/or Spring\n AI<\/b> frameworks for LLM orchestration within Java applications.<\/span><\/span>
<\/li>- Understanding of AI agent concepts \-\n tool/function calling, agent loops, multi\-step task orchestration.<\/span><\/span>
<\/li><\/ul> <\/span><\/span>Good\-to\-Have\nSkills<\/span><\/span><\/b>
<\/p>- Experience with vector databases<\/b> such as\n Pinecone, Weaviate, Qdrant, pgvector, or Chroma for embedding storage and\n semantic search.<\/span><\/span>
<\/li>- Exposure to multiple LLM providers<\/b> (OpenAI, Anthropic Claude, AWS Bedrock, LLaMA/Mistral via Ollama or\n similar).<\/span><\/span>
<\/li>- Familiarity with open\-source LLM deployment<\/b> \- running models locally or on self\-hosted infrastructure.<\/span><\/span>
<\/li>- Knowledge of token budgeting<\/b>, context\n window management, and cost optimization strategies for LLM APIs.<\/span><\/span>
<\/li>- Experience with observability tooling<\/b> for\n AI systems \- LangSmith, Helicone, or custom logging pipelines.<\/span><\/span>
<\/li>- Familiarity with streaming responses<\/b> from\n LLMs (Server\-Sent Events / WebSocket delivery to frontend).<\/span><\/span>
<\/li>- Understanding of embedding models<\/b> \- how\n they differ from generative models, when to use which, and how to evaluate\n embedding quality.<\/span><\/span>
<\/li>- Exposure to CI/CD pipelines<\/b> and\n containerization (Docker, Kubernetes).<\/span><\/span>
<\/li><\/ul> <\/span><\/span>
<\/p>
<\/p>
<\/div><\/span>
