Sandeep Poloju

Software Engineer

United States2 yrs 7 mos experience

AI ML PractitionerAI Enabled

Key Highlights

3+ years of experience in full-stack development.
Expertise in AI infrastructure and machine learning pipelines.
Top 0.06% in national coding competition.

Stackforce AI infers this person is a Full Stack Software Engineer with a focus on AI and Data Engineering in the Fintech sector.

Contact

Skills

Core Skills

JavaMicroservicesMachine LearningData EngineeringFull Stack DevelopmentAi Infrastructure

Other Skills

Java Spring BootJWT-based authenticationRBACAzure Container AppsGitHub Actions CI/CDAWS S3TesseractOCRBeautifulSoupSQLAWS LambdaAPI GatewayD3.jsNeo4jPythonFastAPI

About

Software Engineer with 3+ years of experience building production systems for Fortune 500 clients. Master's in Software Engineering from UMD, College Park (3.93 GPA). I started my career at Infosys on a team that operated more like a startup than a large enterprise. We built a Knowledge Graph platform for financial and insurance clients, the kind of product where complex connected data needed to be made simple enough for analysts to explore visually. I worked across the full stack, from building the Angular frontend and graph visualization layer to developing Python and Java microservices, writing Cypher queries against Neo4j, and deploying everything on Kubernetes. It was fast-paced, client-facing, and gave me a strong sense of ownership over the features I shipped. At UMD, I worked as a Research Assistant building an AgriTech ML pipeline, everything from data ingestion and ETL to model training on AWS SageMaker and serverless deployment. That experience taught me how to work with messy, real-world data and bridge the gap between research and production. What excites me most right now is the intersection of system design and AI infrastructure. I'm passionate about building software that operates at scale, and I've been exploring how LLMs can be integrated into production systems as reliable, orchestrated components. I'm actively building with LangGraph, LangChain, RAG pipelines, MCP (Model Context Protocol), and agentic AI frameworks. Top 100 (Top 0.06%) in HackwithInfy National Coding Competition (167,000+ participants). 600+ algorithmic problems solved. Published NLP research in IJRAR.

Experience

2 yrs 7 mos

Total Experience

2 yrs 7 mos

Average Tenure

Current Experience

Govierates

Software Developer

Feb 2025 – Present · 1 yr 3 mos · Upper Marlboro, MD

Built a middleware REST API layer using Java Spring Boot that acts as a data sync bridge between the GovieRates accounting platform and Project Magnus (a DCAA-compliant timesheet system), eliminating data mismatches across client accounts.
Implemented JWT-based authentication and RBAC, enforcing role-specific access (admin, employee, manager) across API endpoints to ensure data isolation between client accounts.
Automated deployment to Azure Container Apps using GitHub Actions CI/CD pipeline, enabling zero-downtime releases with containerized builds on every merge to main.

Java Spring BootJWT-based authenticationRBACAzure Container AppsGitHub Actions CI/CDJava+1

University of maryland

Software Engineer

Apr 2024 – Jan 2025 · 9 mos · College Park, MD · Remote

Built a full-stack AgriTech web application to predict Brown Patch disease severity for turfgrass researchers, combining data engineering, machine learning, and cloud deployment.
Built an ETL pipeline to ingest, clean, and unify 10GB+ of heterogeneous data from PDFs (TesseractOCR), web scraping (BeautifulSoup), and NOAA weather APIs into AWS S3
Wrote SQL-based data transformations to normalize and join climate, soil, and fungicide datasets
Implemented stepwise regression for feature selection, collaborating with plant science domain experts to validate that algorithm-chosen features were biologically relevant
Improved model accuracy from 74% to 91% through Koppen climate-based regional segmentation and feature engineering
Trained Random Forest and Neural Network models on AWS SageMaker
Deployed serverless inference via AWS Lambda and API Gateway for cost-efficient, real-time prediction

AWS S3TesseractOCRBeautifulSoupSQLAWS LambdaAPI Gateway+2

Infosys

Software Engineer

May 2021 – Dec 2023 · 2 yrs 7 mos

Joined a team that operated like a startup inside a large enterprise, fast-moving, client-facing, and deeply technical. We built an AI-driven Knowledge Graph platform for Fortune 500 financial and insurance clients, making complex connected data explorable through interactive graph visualization, natural language search, and multi-hop entity traversals.
I worked across the full stack, owning features from the React frontend and graph visualization layer (D3.js, Neo4j) to Python and Java backend microservices, and containerized deployments on Kubernetes via Rancher.
Built interactive graph visualization dashboards using D3.js to render financial entity networks for a team of 15 fraud analysts, enabling real-time investigation of suspicious transaction clusters and reducing manual fraud pattern identification effort.
Optimized the search suggestion API for a high-traffic FinTech platform by re-architecting the core Cypher query and implementing Protobuf, reducing p95 latency from ~200ms to ~20ms.
Integrated a React/Node.js application layer with Palantir Foundry ontology APIs to consume curated financial entity datasets, enabling real-time data-driven insights for the fraud detection platform.
Architected and built a FastAPI-based NLU microservice integrating LLMs to boost search accuracy from 85% to 92% on a financial services platform, with response caching for real-time query understanding.
Simplified microservice data serialization to create a minimal API payload, cutting network traffic by 68% (465KB to 148KB).
Engineered a Python data transformation layer to parse GraphQL query formats into Cypher, eliminating an intermediate service, reducing cross-service coupling, and streamlining data flow across Neo4j and Palantir Foundry.
Established RESTful APIs for User Preferences using Node.js and PostgreSQL, enabling persistent storage of user settings and personalized dashboard configurations.