Looking for a job

Nalan T

@nalant

AI engineer building LLM document-processing systems, with human review for uncertain extractions and measurable hallucination rates.

Germany

Message

I build LLM features as real software, not notebooks. My independent document-processing platform uses Node.js and TypeScript to extract structured fields, validate them against business rules, and route uncertain results to human review rather than writing them silently.

I made the platform model-agnostic, with versioned prompts, JSON-schema validation, and an evaluation harness for accuracy, latency, and hallucination rate. I also instrumented its APIs, Kafka workers, and LLM calls so I can trace a document end to end.

At Bank of America, I shipped customer-facing credit card application journeys under production change control. I also consolidated duplicated markup into a shared React component library across 50+ pages, cutting duplication by 40%, and improved load times to under two seconds.

Experience

Work history, roles, and key accomplishments

Frontend Engineer

Bank of America

Jul 2023 - Apr 2025 (1 year 9 months)

Shipped customer-facing credit card application journeys under live production traffic and owned the daily release process. Worked with Line of Business teams to build UI flows and multi-step approval workflows, and reduced markup duplication by 40% through a shared React component library.

React TypeScript JavaScript UI Development Component Libraries Accessibility Compliance Release Management Monitoring API Driven Rendering

Full Stack Engineering Intern

Bank of America

Jan 2023 - Jun 2023 (5 months)

Built and deployed a fraud detection pipeline on SageMaker with Glue ETL jobs, and developed an ASP.NET Core backend using Kafka and Redis with saga orchestration and transactional outbox. Created a multi-step approval workflow engine with SLA timers and role-based access.