Hasanat Jahan

Full-Stack Software Engineer

Distributed Systems · Automation · AI · Machine Learning

I'm a full-stack software engineer who builds automation and machine learning systems for real-world data platforms. I currently work on healthcare claims infrastructure at Smarter Technologies, building distributed services and automation for claims and remittance processing. Previously, I was a Software Development Engineer on Amazon Web Services' Aurora Global Database team, where I designed distributed systems and monitoring for AWS's managed database service.

Outside of engineering, I read, write, and occasionally make art.

Self-portrait illustration by Hasanat Jahan, a hand-drawn ink profile with a yellow outline

Experience

Software Engineer, Platform

Jul 2025 – Present

Smarter Technologies · previously Forward Deployed Engineer

  • Architected Enhanced Claims Status in Python, FastAPI, and Temporal, unifying EDI 277/835 data across Stedi, Availity, Waystar, and payer channels, reaching 90%+ AI disposition accuracy across 15,000+ UAT claims and cutting AR follow-up time from 15–25 minutes to under 5.
  • Built payer-ID crosswalks, claim-ID normalization, and X12 parsing across major payers including UHC, Cigna, and BCBS, improving claim-matching reliability and downstream data consistency.
  • Built AI-assisted migration and code-review tooling using Python AST analysis and Cursor AI, cutting portal development time by 60% and automating migrations across 20+ payer packages.
  • Operated distributed services with OpenTelemetry, AWS, Kubernetes, Docker, and Terraform, supporting 100,000+ monthly claims transactions while reducing transaction exception rates to 4%.

Software Development Engineer

Sep 2022 – Mar 2024

Amazon Web Services · Aurora Global Database

  • Led end-to-end design and implementation of distributed monitoring for Aurora Global Database's flagship feature, improving visibility into system health while maintaining high availability, low latency, and cost efficiency.
  • Streamlined PostgreSQL data flows, implementing change data capture and snapshot handling, and integrated monitoring into the broader Aurora architecture, increasing the reliability and efficiency of data handling.
  • Designed and deployed customer-facing monitoring and alerting (metrics, logs, dashboards, and internal APIs) to detect anomalies quickly and reduce time to resolution for distributed database workloads.
  • Contributed to full-stack development in Java, Python, and C++, participating in design and code reviews and collaborating with front-end teams to ship high-quality, testable features.
  • Built a comprehensive automated testing infrastructure (unit, integration, end-to-end, and CI pipelines) that improved release confidence and reduced regressions, and developed internal tools and troubleshooting workflows that enabled faster root-cause analysis in production.
  • Mentored new hires and interns, creating onboarding resources and pairing on critical systems, which reduced onboarding time and strengthened team performance.

Before that, I was a Data Science Student in Microsoft's Data Science Summer School. I studied Computer Science with a minor in Computational Linguistics at Queens College, graduating Summa Cum Laude with a 3.96 GPA (Phi Beta Kappa).

Skills

Programming & Backend
Python, Java, C++, C, JavaScript, SQL, PHP, R; FastAPI, Pydantic, Node.js, Express.js, React.js, MongoDB, REST APIs, Microservices, Temporal
Cloud & Distributed Systems
AWS (EC2, S3, RDS, Aurora, Lambda, CloudFormation, CloudWatch), Kubernetes, Docker, Terraform/OpenTofu, PostgreSQL, CDC, Distributed Systems
AI & Operations
AI evaluation & UAT, AI-assisted developer tooling, OpenTelemetry, HyperDX, CI/CD, Monitoring, Alerting, Production Troubleshooting
Healthcare
EDI/X12 (835, 837, 277, 271), Claims Status, Remittance Advice, Eligibility, Payer & Clearinghouse Integrations

Projects/Tinkering

Homepage of hasanatjahan.com showing a grid of paintings and sketchbook spreads with their titles and materials

Art Portfolio, hasanatjahan.com

A gallery for my paintings, drawings and sketchbook work, built with Next.js 16 and TypeScript. Deployed on Vercel from main to a custom apex domain, with a GitHub Actions pipeline that gates every merge on lint, typecheck, build-time content validation and a full build.

Continually a work in progress. New work and features land as I make them.

 Unbeatable Tic-Tac-Toe web app on a dark background, with a scoreboard counting human wins, robot wins and draws above an empty board

Unbeatable Tic-Tac-Toe

A browser Tic-Tac-Toe game played against a bot. The Minimax algorithm with alpha-beta pruning, implemented in JavaScript, ensures the human opponent never wins. A shared scoreboard counts everyone's games.

Despair not. The robots have not won, math did. Grew out of an earlier Python command-line version.

 Happy and sad face icons illustrating sentiment analysis

Movie Review Sentiment Classification With Naive Bayes

A Naive Bayes classifier built in Python that classifies IMDB movie review sentiment, reaching 89.712% accuracy with optimizations.

 Output listing word similarity scores from a Word2Vec model, such as 'dirty' and 'clean' at 0.38

Word Similarity with Word2Vec

Trained a Word2Vec model in Python to calculate similarity between words in vector space.

 Diagram illustrating a maximum likelihood estimate language model predicting the next word in a sentence

Maximum Likelihood Estimate Language Modeling

Built and trained generative MLE language models in Python (unigram, bigram, and bigram with add-one smoothing), evaluated through perplexity calculations.

 Scatter plot comparing actual and predicted values from research extending 'Predicting the Present with Google Trends'

Extending "Predicting the Present with Google Trends"

Replicated results from Choi and Varian's research on Motor Vehicle and Parts trends, then extended it to show that a rigorous baseline using public Census data performs almost as well as trends data alone.

 README screenshot of the Documentation Generator tool describing its features and installation steps

Documentation Generator

An automated tool that leverages the GitHub and Greptile APIs to index repositories, generate README files, and streamline the documentation workflow.

 Screenshot of The Dojo Blog, a React blog app listing several blog posts

React Blog App

A simple blog app built with React, featuring basic CRUD functionality, client-side routing, and a JSON server for mock data.

 MERN stack logo graphic representing MongoDB, Express, React, and Node.js

CRUD App Backend

A simple backend API for managing products using Node.js, Express, and MongoDB. Supports basic CRUD operations and JSON data exchange.

 Screenshot of the WikiDiver app with a duck illustration, search box, and a list of Wikipedia article results

Wikipedia Searcher

Uses JavaScript and jQuery against the MediaWiki API to search Wikipedia articles from a search box, or pull up a random article with one click.

Quack the duck for an interesting surprise.

 Weather app screenshot showing a rain icon, 16°C temperature, and Queens County, US location

Online Weather Viewer

Displays the weather for the user's current location from a JSON API response, with a temperature toggle and weather icon.

 Shiba Inu dog photo above a button to post the image to Imgur

Imgur Shibe Poster

A web application that authenticates with the Imgur API and fetches random Shiba Inu images from the shibe.online API, posting them directly to a user's Imgur account with automatic token caching.