Senior Software Engineer (AI/ML)
SoftServe
● Designed and delivered LLM-powered features for enterprise clients in retail, healthcare, and SaaS, covering document intelligence, conversational assistants, and content generation workflows. ● Built production RAG pipelines using LangChain, OpenAI APIs, and Pinecone, improving answer relevance on internal knowledge bases and reducing manual document lookup time by roughly 40%. ● Developed and fine-tuned ML models in PyTorch and scikit-learn for recommendation, classification, and demand forecasting use cases on large client datasets. ● Architected scalable inference services with FastAPI, Celery, and Redis, supporting high-volume API traffic with stable latency under load. ● Reduced model inference costs by approximately 30% through prompt optimization, response caching, batching, and right-sizing of model selection per use case. ● Implemented MLOps workflows with MLflow, Docker, Kubernetes, and GitHub Actions, cutting model deployment turnaround from days to hours. ● Built evaluation frameworks for LLM outputs covering accuracy, hallucination checks, and regression testing before each release. ● Integrated AI capabilities into existing client platforms through REST APIs, webhooks, and event-driven microservices on AWS. ● Developed React.js and TypeScript interfaces for AI tooling, including admin consoles, analytics dashboards, and human-in-the-loop review screens. ● Designed PostgreSQL and pgvector schemas for embeddings storage, semantic search, and audit logging of AI interactions. ● Implemented secure access patterns using OAuth 2.0, JWT, and role-based permissions across AI service endpoints. ● Partnered with product managers, data scientists, and client stakeholders to scope AI roadmaps and translate requirements into delivery milestones. ● Mentored mid-level engineers on LLM application patterns, code quality, and testing practices through reviews and pairing sessions.