Looks like you have JavaScript disabled. For the full Twine experience, you will need to re-enable it.

Full stack software engineer (7+ years) shipping production AI systems end to end—LLM-powered document extraction pipelines, agentic workflows, and the backends and interfaces around them. AWS certified and comfortable prototyping quickly with LLMs, then hardening for real users. Currently architecting an AI-assisted system of record that replaces spreadsheet-driven enterprise workflows using prompt engineering with strict schema-validated JSON output, RAG, provenance tracking, and human-in-the-loop validation—ensuring every AI-generated value is traceable and reviewable.…Full stack software engineer (7+ years) shipping production AI systems end to end—LLM-powered document extraction pipelines, agentic workflows, and the backends and interfaces around them. AWS certified and comfortable prototyping quickly with LLMs, then hardening for real users. Currently architecting an AI-assisted system of record that replaces spreadsheet-driven enterprise workflows using prompt engineering with strict schema-validated JSON output, RAG, provenance tracking, and human-in-the-loop validation—ensuring every AI-generated value is traceable and reviewable.

Yashamrutiya

Full Stack Developer, AI Engineer, Software Engineer, +3





Full stack software engineer (7+ years) shipping production AI systems end to end—LLM-powered document extraction pipelines, agentic workflows, and the backends and interfaces around them. AWS certified and comfortable prototyping quickly with LLMs, then hardening for real users. Currently architecting an AI-assisted system of record that replaces spreadsheet-driven enterprise workflows using prompt engineering with strict schema-validated JSON output, RAG, provenance tracking, and human-in-the-loop validation—ensuring every AI-generated value is traceable and reviewable.…Full stack software engineer (7+ years) shipping production AI systems end to end—LLM-powered document extraction pipelines, agentic workflows, and the backends and interfaces around them. AWS certified and comfortable prototyping quickly with LLMs, then hardening for real users. Currently architecting an AI-assisted system of record that replaces spreadsheet-driven enterprise workflows using prompt engineering with strict schema-validated JSON output, RAG, provenance tracking, and human-in-the-loop validation—ensuring every AI-generated value is traceable and reviewable.

Available to hire

Full stack software engineer (7+ years) shipping production AI systems end to end—LLM-powered document extraction pipelines, agentic workflows, and the backends and interfaces around them. AWS certified and comfortable prototyping quickly with LLMs, then hardening for real users.

Currently architecting an AI-assisted system of record that replaces spreadsheet-driven enterprise workflows using prompt engineering with strict schema-validated JSON output, RAG, provenance tracking, and human-in-the-loop validation—ensuring every AI-generated value is traceable and reviewable.

Skills

Experience Level

Expert

Expert

Expert

Expert

Expert

Expert

Expert

Expert

Expert

Expert

Expert

Expert

Expert

Expert

Expert

Intermediate

Intermediate

Intermediate

Language

Work Experience

Software Engineer — AI Document Intelligence Platform (Independent, Remote) at Independent (Toronto, ON)

March 1, 2026 - Present

Architecting an AI-assisted system of record for the Canadian group benefits insurance industry to replace Excel-based broker renewal workflows with LLM-powered extraction, a canonical data schema, and human-in-the-loop validation. Engineered LLM extraction prompts producing strict schema-validated JSON with per-field provenance, enabling auditability by tracing AI-generated values to source page(s). Built document ingestion pipelines that classify and split multi-record scanned folders (barcode separators, mixed carrier formats) into structured, individually indexed records. Designed a canonical data model with 16+ entities spanning policies, benefit lines, rates, and validation results, including reverse-engineering a 107-sheet broker workbook and surfacing a production formula defect in the incumbent process. Operate the full stack solo using MongoDB and Dockerized services on Windows Server, including compliance-oriented backup strategy and agentic AI developer tooling to increase

Software Engineer — AI Document Intelligence Platform at Independent (Toronto, ON) (Remote)

March 1, 2026 - Present

Architecting an AI-assisted system of record for the Canadian group benefits insurance industry, replacing Excel-based broker renewal workflows with LLM-powered extraction, a canonical data schema, and human-in-the-loop validation. Engineered extraction prompts to produce strict schema-validated JSON with per-field provenance, enabling traceability of AI-generated values to source pages for audit and review. Built ingestion pipelines to classify and split multi-record scanned folders into structured, individually indexed records. Designed a canonical data model spanning policies, benefit lines, rates, and validation results by reverse-engineering a 107-sheet production broker workbook and surfacing a live formula defect in the incumbent process. Operate the full stack solo including a MongoDB backend, Dockerized services on Windows Server, compliance-oriented backup strategy, and agentic AI developer tooling.

Software Developer at Toshiba Tec Canada Business Solutions Inc.

June 1, 2021 - February 1, 2026

Produced AI applications extracting key information from business documents using Google Document AI, improving extraction accuracy by 40%. Engineered integrations connecting Salesforce, QuickBooks, and e-signature platforms with M-Files, expanding cross-platform functionality and increasing revenue by 10%. Created customized invoice-processing and document-management workflows for enterprise clients, cutting processing time by 40%. Designed applications for Toshiba Multifunction Printers enabling fax/document management from the device interface, improving processing speed by 30%. Developed Progressive Web Apps with tailored on-site data collection forms to improve adoption by 25% and reduced system downtime by 50% via proactive debugging and performance optimization.

Associate Software Engineer at Arastu Systems, Inc.

January 1, 2018 - July 1, 2019

Built single-page applications and real-time high-transaction systems on the MEAN stack (MongoDB, Express, Angular, Node.js).