A
Ashutosh Gupta
About
Detail
California, United States
Résumé
Jobs
verified_user
0% verified
AI Backend EngineerUnitedHealth GroupJan 2025 - Current (1 year 8 months)As an AI Backend Engineer, I led the development of a clinical-summary API in Python (FastAPI) integrated with a Retrieval-Augmented Generation (RAG) pipeline using LangChain and FAISS, which efficiently retrieves CPT/ICD guidelines and historical claims, resulting in a 40% reduction in prior-auth turnaround time. I designed an AI-driven claim-triage engine employing large language model (LLM) reasoning agents to assess coding consistency, medical necessity, and duplicate intent, significantly improving auto-decision accuracy from 67% to 86%. Additionally, I developed a natural language processing (NLP) layer utilizing spaCy and Hugging Face Transformers to extract diagnosis, drug, and procedure concepts from physician notes and appeal file- Software EngineerHhaexchangeDec 2020 - Jul 2023 (2 years 8 months)As a Software Engineer, I designed multi-tenant visit and authorization services in Java (Spring Boot) backed by Aurora PostgreSQL, adding sharded lookup tables that lifted read throughput by 42% for high-volume agencies. I modeled scheduling and caregiver-management UI in React with selective prefetching and server-driven pagination, successfully cutting screen load time from approximately 3 seconds to under 900 ms. Additionally, I implemented real-time visit event streams using Kafka to sync EVV punches, shift edits, and plan-of-care changes, which reduced stale-visit mismatches by 31%. I created batch payroll and billing processors using Spring Batch and S3 to validate Medicaid file formats, provider rates, and rounding policies, ensurin
- SProduct Insight AnalystSage SofttechFeb 2018 - Nov 2020 (2 years 10 months)As a Product Insight Analyst, I architected multi-tenant subscription and billing APIs using Python (Django) and PostgreSQL, which resulted in a 23% reduction in invoice and proration errors across customer accounts. I structured background workers in Celery and Redis to efficiently handle usage events, emails, and webhooks, cutting notification and sync delays by nearly 80% on peak days. Additionally, I assembled weekly product metrics pipelines in Python (Pandas) from raw clickstream and feature-usage logs, enabling the team to run experiments that lowered churn by approximately 12% in two key segments. I tuned database performance through targeted PostgreSQL indexing and query refactors, improving the 95th-percentile dashboard load time
Education
verified_user
0% verified
- Master of Science in Computer ScienceIndiana University BloomingtonAug 2023 - May 2025 (1 year 10 months)
Awards
verified_user
0% verified
- HBest EmployeeHHA ExchangeJun 2022 - Jun 2023 (1 year 1 month)