Nova Web Technologies

Compensation

Full Description

Location: Remote Experience: 8+ years (senior-level) Employment Type: Full-time / Contract Position overview: You will design and implement resilient, secure, and observable backend microservices in Python and Node.js that integrate with AI platforms (AWS Bedrock and others), knowledge bases, vector stores, and LLM orchestration frameworks. You’ll own API design, data flows for embeddings & retrieval, agent task orchestration, and production readiness (CI/CD, monitoring, security). Key responsibilities: • Design, build, and maintain backend microservices and APIs in Python and Node.js (REST/GraphQL/gRPC). • Implement AI-integration layers: embedding generation, vectorization, vector search, knowledge-base connectors, and RAG pipelines. • Integrate with LLM & model-hosting platforms (AWS Bedrock and comparable offerings), and implement prompt chaining and agent orchestration. • Build, test, and operate AI agent frameworks that enable multi-step, stateful workflows and task delegation. • Architect event-driven and serverless components to scale real-time inference and asynchronous jobs. • Ensure services are production-ready: automated CI/CD, Infrastructure-as-Code, logging, tracing, and alerting. • Collaborate with data scientists, ML engineers, product owners, and security teams to productionize models and guardrails (content filtering, rate limits, auditing). • Optimize for cost, latency, and throughput in model calls, embedding pipelines, and vector-index maintenance. • Produce clear API docs, runbooks, and design docs; participate in architectural reviews. Required Skills: • Strong backend experience (5+ years) building microservices using Python and/or Node.js. • Hands-on with AI integrations: building APIs that call LLMs, generate embeddings, and perform vector searches. • Experience implementing RAG pipelines, knowledge-base connectors, and vectorization workflows. • Familiar with LLM orchestration techniques: prompt chaining, agent frameworks, multi-step reasoning flows, and action/tool invocation. • Practical experience integrating with AWS Bedrock or other managed LLM offerings (production-level integrations). • Solid engineering grounding in API design, authentication/authorization, retry/backoff patterns, and rate-limiting. • Familiarity with at least one vector store (e.g., Pinecone, Milvus, Weaviate, or vector DBs on AWS/Aurora with pgvector) and retrieval techniques. • Track record of shipping production services with CI/CD, automated tests, and monitoring. Preferred qualifications: • Bachelor’s or Master’s degree in Computer Science, Engineering, or equivalent practical experience. • Prior experience building AI-powered chatbots, virtual assistants, or automated agents in production. • Experience with secure handling of PII and building data retention / redaction pipelines. • Familiarity with model cost-optimizations (batching, caching, hybrid retrieval) and legal/regulatory considerations for LLM usage. Nice to have: • Deep knowledge of cloud-native architecture and event-driven systems. • Comfortable with AWS: Lambda, DynamoDB, S3, API Gateway, ECS/EKS, and IAM best practices. • Experience with embedding-generation pipelines, approximate nearest neighbor (ANN) algorithms, vector index tuning, and vector-store maintenance. • Familiarity with LLM orchestration libraries / agent frameworks (e.g., LangChain, LlamaIndex, Agents paradigms) and prompt engineering best practices. • Domain-Driven Design (DDD) and strong modeling skills for service boundaries. • Experience with observability stacks (CloudWatch, OpenTelemetry, Prometheus, Grafana, ELK). • Comfortable with Git-based collaborative workflows, feature branching, code reviews, and trunk-based development. • Containerization (Docker) and Kubernetes or serverless deployment experience.