IS
Full Stack Developer - Chatbot Platform
ITCO Solutions
- 🇺🇸 United States
- Remote
- Mid level
- 5 hours ago
- AI
- Machine Learning
- RAG
- Python
- Vector Search
- LLM APIs
- LangChain
- LangGraph
- Semantic Kernel
- LlamaIndex
- SQL
- AWS
- Amazon Bedrock
- Azure OpenAI
- Snowflake
- Redshift
- BigQuery
- OpenSearch
- Pinecone
- pgvector
- ETL
- AWS Glue
- Azure Data Factory
- CI/CD
- Docker
- ECS
- Kubernetes
- GitHub Actions
5 hours ago
Location: Burlington preferred, but remote is acceptable
Client candidates accepted
This is an urgent need and includes candidate interview availability.
Position Description
Our client is seeking an experiencedAI Chatbot Developer to design, build, deploy, and support production grade generative AI and conversational AI solutions. This role will focus on developing reliable LLM powered applications, including Retrieval Augmented Generation (RAG), prompt and orchestration workflows, AI safety controls, evaluation frameworks, and integrations with enterprise data and applications.
The successful candidate will bring strong Python development skills, practical experience implementing LLM applications in production environments, and the ability to troubleshoot across prompts, models, retrieval pipelines, APIs, data sources, and application code.
Your future duties and responsibilities
- Design, develop, and maintain production grade LLM and conversational AI applications using Python and modern API based architectures
- Build and optimize Retrieval Augmented Generation solutions, including document ingestion, chunking, embeddings, retrieval strategies, vector search, context construction, and retrieval relevance
- Design, test, version, and improve prompts, system instructions, grounding strategies, structured outputs, and multi step LLM interactions
- Implement tool and function calling, conversation context management, agentic workflows, and deterministic orchestration patterns based on application requirements
- Develop evaluation frameworks, representative test sets, regression testing processes, and measurable quality criteria for LLM powered applications
- Implement AI safety controls, including prompt injection defenses, content filtering, grounding mechanisms, input and output validation, and safeguards against unsupported responses
- Implement and support observability and traceability capabilities capturing prompts, retrieved context, model responses, tool calls, errors, latency, and other production telemetry
- Investigate and resolve issues across the full LLM application stack, including prompts, retrieval systems, models, APIs, enterprise data sources, and application code
- Collaborate with business and technical stakeholders to translate ambiguous requirements into reliable AI behaviors, technical solutions, and measurable acceptance criteria
- Develop and support secure cloud based AI solutions while protecting sensitive data, managing credentials and access controls, and ensuring data remains within approved environments
- Python: Strong Python skills for backend services, including API integration, error handling, data serialization, concurrency/asynchronous processing, logging, and maintainable production code.
- LLM application development:Hands on experience building production applications using LLM APIs, including prompt design, tool/function calling, structured outputs, conversation context/history, and orchestration of multi step interactions.
- Prompt engineering:Experience designing, testing, versioning, and improving prompts for production use, including system prompts, grounding instructions, response formatting, and techniques for reducing hallucinations.
- Retrieval Augmented Generation (RAG): Practical experience designing and implementing RAG solutions, including document ingestion, chunking, embeddings, retrieval strategies, vector search, context construction, and improving retrieval relevance.
- LLM evaluation and quality assurance:Experience creating representative test sets, defining evaluation criteria, measuring response quality, performing regression testing, and diagnosing failures to their root cause. Familiarity with automated LLM evaluation tools is strongly preferred.
- AI safety and guardrails:Experience implementing safety controls for LLM applications, including content filtering, prompt injection defenses, grounding controls, input/output validation, and other techniques for limiting inappropriate or unsupported responses.
- Agent and orchestration frameworks: Experience with LLM application or agent frameworks such as LangChain, LangGraph, Semantic Kernel, LlamaIndex, or equivalent frameworks, including understanding when agentic patterns are appropriate versus simpler deterministic workflows.
- SQL and data analysis:Strong analytical SQL skills and the ability to assess whether generated queries and results accurately answer the underlying business question.
- Traceability and observability: Experience implementing logging and traceability for LLM applications, including capturing prompts, retrieved context, model responses, tool calls, errors, latency, and other information needed to troubleshoot and evaluate production behavior.
- Cloud: Practical experience designing, deploying, and operating applications on a major cloud platform, preferably AWS, including familiarity with managed AI/LLM services and cloud security practices.
- Security and data protection:Understanding of secure LLM application design, including protecting sensitive data, controlling access to data sources, managing credentials/secrets, and ensuring that application data is used only within approved environments.
- Production troubleshooting:Ability to investigate and resolve issues across the full LLM application stack, including prompts, retrieval, models, APIs, data sources, and application code.
- Text to SQL, conversational analytics, natural language querying, or BI tools.
- Experience with cloud hosted LLM platforms such as Amazon Bedrock, Azure OpenAI, or comparable services.
- A cloud data warehouse such as Snowflake, Redshift, or BigQuery.
- Vector search technologies such as OpenSearch, Pinecone, pgvector, or equivalent platforms.
- Experience improving RAG quality through techniques such as hybrid search, reranking, metadata filtering, query rewriting, or retrieval evaluation.
- LLM performance, latency, and cost optimization, including caching, model selection, token management, and prompt optimization.
- Data pipelines or ETL technologies such as Spark, AWS Glue, Azure Data Factory, or equivalent tools.
- Containers and CI/CD, such as Docker, ECS, Kubernetes, GitHub Actions, or comparable technologies.
- Monitoring and analytics for production AI applications, including usage, response quality, user feedback, latency, failures, and model/token consumption.
- Experience integrating LLM capabilities into an existing web application or enterprise application rather than building only standalone AI prototypes.
- Education, government, or other public sector data experience.
- Experience working with business users to translate ambiguous questions and requirements into reliable AI behaviors and measurable acceptance criteria.
- Client facing or consulting experience, including the ability to explain AI behavior, limitations, and technical tradeoffs to non technical stakeholders.
Full Stack Developer - Chatbot Platform · ITCO Solutions