ajay@lafayette:~$ whoami

Ajay Banstola

Software Engineer @ Apex Innovations · Lafayette, LA

ajay@lafayette:~$ cat interests.txt

AI systems · fuzzing & binary analysis · medical-imaging ML

ajay@lafayette:~$

psst — this site has a real terminal. Press ` or Ctrl K anywhere.

01 / about

The short version

Software engineer with 5+ years across AI platforms, large-scale archiving systems, and fintech. Currently building at Apex Innovations. I recently finished my M.S. in Computer Science at UL Lafayette with a 4.0, where my research spanned glaucoma progression prediction from RNFLT maps and binary-level fuzzing under Prof. Arun Lakhotia.

02 / skills

What I work with

Languages: Python · Java · JavaScript / TypeScript · C / C++ · Rust (research)
AI / ML: PyTorch · TensorFlow · scikit-learn · CNNs · RAG · LangChain · Hugging Face
Web & Backend: Django · Flask · React · Node.js · REST APIs
Infra & Tools: AWS · Azure · Datadog · Jenkins CI/CD · Git · Linux
Data: PostgreSQL · MSSQL · MongoDB · ETL pipelines · Tableau · Power BI

03 / work

Experience

Software Engineer · Apex Innovations

Mar 2026 — Present

Lafayette, LA

Building software at Apex Innovations. (Edit src/data.ts to describe your current work.)

—

AI Engineer · Opportunity Machine

May 2025 — Aug 2025

Lafayette, LA

Co-led full-stack development of the Blue Partner AI platform with law enforcement partners, including a real-time notification system sharing live map locations with officers — cutting paperwork by 75%.
Integrated GPT for conversational AI, automated report generation, and policy compliance, reducing average incident-report time from 50 to 15 minutes.

AzureReactTypeScriptElevenLabs

Software Engineer · Smarsh Inc.

Nov 2021 — Nov 2023

Portland, OR

Built and deployed a scalable web-archiving platform; automated CI/CD with Jenkins (75% faster deploy cycles), kept 95%+ unit-test coverage, and wired Datadog monitoring across 65+ distributed servers.
Designed high-performance migration pipelines moving multi-terabyte datasets from PostgreSQL to MSSQL, improving transfer efficiency by 20%.

PythonDjangoJenkinsDatadog

Associate Software Engineer · PayNep Pvt. Ltd.

Nov 2020 — Oct 2021

Kathmandu, Nepal

Led back-end and API development for a digital wallet — bus ticketing, mobile top-up, and payment integrations — and led its deployment.
Documented APIs in Confluence and automated Postman test runners, improving API reliability for cross-functional teams.

PythonFlaskPostman

04 / projects & research

Things I've built and studied

Glaucoma Progression Prediction

Deep-learning pipeline detecting glaucoma progression from RNFLT maps (GDP500 dataset). Benchmarked VGG and ResNet with subgroup AUCs across demographic groups to check fairness of generalization.

PyTorchVGGResNet

Byteri — binary-level fuzzing

Enhanced a taint-analysis framework under Prof. Arun Lakhotia: built a custom AFL++ mutator using Byteri's AST-based input generation, and compared taint behavior between C and Rust binaries.

CRustAFL++LLVM

MedScan — pneumonia detection

Web app detecting pneumonia from chest X-rays: CNN model served through a Flask API with a React frontend for real-time predictions.

FlaskCNNReact

Customer Predictive Analytics

Lead-conversion prediction with logistic regression on historical customer data, with a web dashboard visualizing the results.

scikit-learnFlask

05 / journey

Kathmandu → Portland → Lafayette

2026 — now current
Software Engineer

Apex Innovations, Lafayette LA
2025 graduated
M.S. Computer Science, 4.0 GPA

UL Lafayette — Academic Excellence Award
2025 ai
AI Engineer

Opportunity Machine — Blue Partner platform
2024 fellowship
Robert May Fellowship

Full-duration fellowship for the M.S. program
2021 — 2023 scale
Software Engineer

Smarsh Inc., Portland OR
2020 — 2021 start
Associate Software Engineer

PayNep, Kathmandu — where it started

06 / contact

Say hello

Whether it's AI systems, fuzzing, or a good conversation — always open.