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

I am an AI Software Engineer with more than five years of experience building scalable backend systems, distributed applications, cloud-native services, and AI-powered products. My background includes Python, Java, Django, Spring Boot, FastAPI, microservices, event-driven architecture, Generative AI, LLM-powered autonomous systems, Retrieval-Augmented Generation, and LangGraph.…

Ashish Shah

Software Engineer, AI Engineer, AI Developer, +2





I am an AI Software Engineer with more than five years of experience building scalable backend systems, distributed applications, cloud-native services, and AI-powered products. My background includes Python, Java, Django, Spring Boot, FastAPI, microservices, event-driven architecture, Generative AI, LLM-powered autonomous systems, Retrieval-Augmented Generation, and LangGraph.…

Available to hire

I am an AI Software Engineer with more than five years of experience building scalable backend systems, distributed applications, cloud-native services, and AI-powered products. My background includes Python, Java, Django, Spring Boot, FastAPI, microservices, event-driven architecture, Generative AI, LLM-powered autonomous systems, Retrieval-Augmented Generation, and LangGraph.

Skills

Experience Level

Expert

Expert

Expert

Expert

Expert

Expert

Expert

Expert

Expert

Expert

Expert

Expert

Expert

Expert

Expert

Expert

Expert

Expert

Expert

Expert

Expert

Expert

Expert

Expert

Expert

Intermediate

Intermediate

Intermediate

Intermediate

Intermediate

Intermediate

Intermediate

Intermediate

Work Experience

Backend Software Engineer (AI/ML) at Uber

June 1, 2026 - Present

Developed LangGraph-based query preprocessing and source-identification agents for an internal AI chat platform serving more than 400 engineers. Architected Python and FastAPI orchestration and inference services for a multi-stage RAG pipeline, reducing response time by four seconds and search space by 70%. Built an LLM-as-a-judge evaluation harness that reduced SME review time from 12 minutes to under three minutes per response. Automated Docker and Kubernetes deployment workflows across 12 AWS production services, reducing release windows from four hours to under one hour.

Software Engineer (Volunteership) at The University of Texas at Arlington

August 1, 2025 - June 1, 2026

Spearheaded a multi-agent diabetes care system using LangGraph and LangChain with four ReAct agents, tool integration, conversational memory, and Amazon Bedrock inference. Built an agentic RAG pipeline over ADA, WHO, and ICMR clinical guidelines using a FAISS vector store and an LLM-based domain guardrail that filtered over 95% of unsafe or out-of-scope requests. Containerized components with Docker on AWS for six concurrent research environments. Also developed a Django, PostgreSQL, and React financial reconciliation platform that reduced manual effort by 70%.

Software Engineer at LTIMindtree

October 1, 2020 - July 1, 2023

Built event-driven microservices for Citi's Global Accounting Engine using Java, Spring Boot, and Apache Kafka, supporting throughput above 5,000 events per second. Implemented idempotent journal writes, Dead Letter Queues, and cooperative-sticky partition assignment, reducing processing latency from 5,000 to 3,500 milliseconds. Developed a Django and PostgreSQL analytics application for more than 150 users, with REST APIs, JWT authentication, RBAC, asynchronous reporting using Celery and Redis, and secure S3 file delivery. Delivered React dashboards and expanded automated testing across seven backend microservices.

Software Developer at Dell Technologies

January 1, 2019 - September 1, 2020

Developed Java Spring Boot and Hibernate backend services for asset management applications used by more than 2,000 internal users. Refactored 15 legacy backend modules into shared service components, removing more than 6,000 redundant lines of code. Configured GitHub Actions and Jenkins CI/CD pipelines that reduced deployment time from four hours to under one hour per release. Optimized Oracle SQL queries and stored procedures across 25 database routines, reducing inventory job execution times by 40%. Resolved more than 20 production issues through root cause analysis.