Headline

Software Engineer | AI Systems Engineer | RAG at Scale | Cloud-Native AI Infrastructure

About

Backend Engineer with 3.5+ years of production experience, now building AI-powered systems at scale. At ConnxAI Inc, I architected RAG-based diagnostic pipelines serving 9,000+ retail locations, engineered real-time AI platforms on AWS ECS with sub-2-second latency, and optimized automation execution by 88% through async redesign. Core expertise: LLM integration, vector search (FAISS, Pinecone), agentic workflows, cloud-native infrastructure (AWS, Kubernetes, Docker, Terraform), and backend systems (FastAPI, Django, PostgreSQL). Previously at Infor Global Solutions, delivering enterprise SaaS modules and AWS-integrated CI/CD pipelines. Transitioning deeper into core AI engineering model development, fine-tuning, and agentic systems. Open to Backend Engineer (AI team) and AI Systems Engineer roles.

Location

Hyderabad

Work experience

Software Engineer at ConnX Inc., Jul 2024 - Present · 1 yr 11 mos, Hyderabad, Telangana, India, AI Network Troubleshooting System - RAG at 9,000+ Retail Locations Architected and deployed an AI-powered diagnostic system using Retrieval-Augmented Generation (RAG) with FAISS vector search and LLM reasoning. On store outage detection, the system fetches live diagnostics from MIST AI and ServiceNow, queries a proprietary knowledge base, and surfaces resolution steps to operators in real time. Scale: 9,000+ retail locations Tech stack: LangChain, OpenAI API, FAISS, FastAPI, AWS ECS, Docker Impact: Reduced mean time to resolution (MTTR) for network outages, enabling self-service troubleshooting for field operators Real-Time Contact Center Monitoring Platform Engineered a centralized RESTful API layer consuming Webex CC API data Performance: Sub-2-second latency for real-time agent monitoring Tech stack: FastAPI, AWS ECS, Docker, PostgreSQL, WebSockets Impact: Eliminated manual dashboard checks, enabling proactive contact center management Network Automation Optimization: Overhauled legacy automation scripts with an async-first architecture using ThreadPoolExecutor (20 workers). Performance: 88% reduction in execution time Impact: Accelerated configuration rollouts and firmware updates across the entire fleet

Associate Software Engineer at Infor, Aug 2022 - Nov 2023 · 1 yr 4 mos, Hyderabad, Telangana, India

Software Development Intern at Infor, Apr 2022 - Aug 2022 · 5 mos, On-site

Education

Bachelor of Technology - BTech, Computer Science, sree vidyanikethan engineering college

Public links

github: https://github.com/marshmello765