Last login: Thu Oct 08 2026

Man typing
reghu@portfolio:~$ whoami

Reghunaath Ajith Kumar Ahila

reghu@portfolio:~$ cat role.txt

I build full-stack products end-to-end and ship fast.

MS Data Science @ Northeastern · 4x Hackathon Winner · Published Researcher.

> ▋
reghu@portfolio:~$ ls
reghu@portfolio:~$ ./portfolio-quest.exe
reghu@portfolio:~$ ls -la projects/
$cat rescueline-ai/README.md
# RescueLine AI
🏆 1st Place — $700 · Innovaite Hackathon

AI powered emergency call triage system using Twilio, ElevenLabs, and FastAPI to automatically classify and route emergency calls by urgency during natural disasters, when traditional helplines are overwhelmed. Built a real time voice AI agent and live dashboard with WebSocket based updates for emergency coordinators managing high call volumes.

[FastAPI][Twilio][ElevenLabs][Python][WebSockets]
github/backendgithub/frontend
$cat doodlpop/README.md
# Doodlpop
🏆 1st Place · SharkHack

AI powered comic book generator that turns a single sentence into a fully illustrated comic. Type a story idea, Doodlpop generates a panel by panel script with dialogue and visual descriptions, lets you pick an art style (manga, western, watercolor storybook), edit the script, then illustrates every panel with AI. Supports shareable links, QR code sharing, and PDF export.

[Next.js][TypeScript][Gemini][Vercel][AI Image Gen]
githublive demo
$cat leadcatch-ai/README.md
# LeadCatch AI
🥈 2nd Place — $1,500 · Yconic AI Hackathon

Automated SMS based lead conversion powered by OpenAI and Twilio APIs, with a scalable Python backend for multi user handling. After a small business or a trade worker misses a call, an AI agent texts the caller back within seconds, answers their questions, qualifies the lead, and books an appointment straight into the owner's schedule. Turns missed calls, which would otherwise walk away as lost revenue, into booked jobs without the owner ever having to pick up the phone.

[Python][OpenAI][Twilio][SMS][FastAPI]
github
$cat snapback/README.md
# SNAPBACK
🥈 2nd Place · Babson Generator Build-a-thon

Computer vision tool that measures athletic mobility loss after injury or a long break, no wearables, no clinic visit. Pick your sport, stand in front of your camera, and get a mobility score out of 100 benchmarked against clinical reference ranges. MediaPipe and OpenCV track 33 skeletal landmarks at 30fps and compute joint angles in real time. Outputs a sport specific gap analysis and a personalised week by week return to sport exercise plan with sets, reps, and reasoning.

[Python][MediaPipe][OpenCV][Computer Vision][AI]
github
$cat deadpool/README.md
# DEADPOOL

DEADPOOL (Dependency Evaluation And Downstream Prediction Of Operational Liabilities), a multi agent AI system that identifies operational risks and failure cascades in startups before they become catastrophic. Orchestrated 6 specialist agents (People, Finance, Infrastructure, Product, Legal, Code Audit) running concurrently via LangGraph. Cross domain cascade expansion using Gemini 2.5 Pro traces how a single failure propagates across the org, producing a 0–100 composite risk score and a plain language founder briefing.

[LangGraph][Gemini 2.5 Pro][FastAPI][React][Python][Multi-Agent]
github
$cat research/ajith2021activity.pdf
@ 1D Convolution approach to human activity recognition using sensor data and comparison with machine learning algorithms
★ 63 citations

Classifies six human activities (sitting, standing, walking, stair climbing, laying down) from smartphone accelerometer and gyroscope data. Benchmarks classic ML algorithms — Logistic Regression, Linear & Kernel SVM, Decision Tree, Random Forest — against a proposed feed-forward DNN and 1D CNN. Evaluated on Recall, Precision, and F1 score; SVM and the proposed 1D CNN emerged as the best-performing models.

International Journal of Cognitive Computing in Engineering, 2021

[research][deep learning][IoT]
reghu@portfolio:~$ history --jobs
QuantUniversity—Graduate Intern
Boston, MA · Jul 2025 – Aug 2025 · 2 mos
  • >Built a platform using React and FastAPI to enable AI-assisted educational content creation and seamless hosting of generated materials, reducing content development time from 5 days to ~3 hours.
  • >Identified and fixed a critical paywall bypass vulnerability in the first week, strengthening platform security.
  • >Designed and implemented authentication and authorization systems addressing all security and compliance requirements for ISO and SOC2 certification.
[React][FastAPI][Python][Auth]
Northeastern University—Teaching Assistant
Remote · May 2025 – Dec 2025 · 8 mos
  • >Served as TA for CS6620 Fundamentals of Cloud Computing and DS3000 Foundations of Data Science, supporting a combined cohort of graduate and undergraduate students.
  • >Assisted in grading, designing assignments, and holding office hours to clarify concepts and guide students through course material.
[Cloud Computing][Data Science]
Infosys—Senior Systems Engineer
Bengaluru, India · Aug 2023 – Jul 2024 · 1 yr
  • >Developed and deployed a full-stack application with a .NET microservice architecture and React.js frontend, modernizing a legacy insurance platform through REST APIs, JWT-based authentication, and Redux state management.
  • >Built a .NET rule-based recommendation engine integrating 9 external systems through gRPC, SOAP, and REST APIs, with SQL caching for efficient rule evaluation and policy recommendations.
  • >Developed a Python script to auto-generate unit test cases and Postman integration test cases from business-owned Excel sheets, saving over 65 hours of manual effort long term.
[.NET][React][Redux][gRPC][SQL][Python]
Danske IT—Associate Software Engineer
Bengaluru, India · Jul 2022 – Aug 2023 · 1 yr 2 mos
  • >Developed and integrated Camunda BPM workflows within the .NET backend to orchestrate customer onboarding processes, improving data processing efficiency, fault tolerance, and system reliability.
  • >Built and owned CI/CD pipelines on Azure DevOps, ensuring smooth deployment workflows and continuous integration across development and production environments.
  • >Integrated automated load testing with Grafana K6 into the CD pipeline to evaluate system performance and ensure scalability under high traffic.
  • >Independently implemented a monitoring solution using Kibana (Elastic Stack) to provide real-time insights across multiple team projects.
[Camunda BPM][.NET][Azure DevOps][K6][Kibana]
Danske IT—Apprentice
Bengaluru, India · Jan 2022 – Jul 2022 · 7 mos
  • >Gained comprehensive experience in fintech software development, working across testing, DevOps, frontend, and backend in an agile environment.
  • >Improved unit test line coverage from 60% to 95% for .NET backend.
[.NET][Unit Testing]
reghu@portfolio:~$ cat ~/.education
Northeastern University
Northeastern UniversityGPA: 3.9/4.0
MS Data ScienceBoston, MA · 2024 – 2026 · 2 yrs
VIT Vellore
VIT VelloreGPA: 3.42/4.0
B.Tech Computer Science & EngineeringVellore, India · 2018 – 2022 · 4 yrs
reghu@portfolio:~$ cat contact.json
contact.json
reghu@portfolio:~$ echo $UNIQUE_VISITORS