armaan_Day 47 / 84
run my first marathon
18 miles this morning. Longest ever. Legs are toast, brain is quiet for the first time in weeks.
Applaud
Comment
Share
Adopt
David Denny
Backend systems by day. Building Strive by night.
About
01 — About me
Hey, I’m David.
I’m a software engineer. Outside of work, you’ll catch me at the gym, on a pickleball court, or playing with my dog.
— Currently
- Building Strive — preparing for first release
- Wrapping up my M.S. in CS at UT Austin
- Always keeping up with the latest in AI — whether it's new model releases, tools for automation and productivity, or figuring out which models excel at different tasks. I make it a priority to research and experiment so I can stay on top of the ecosystem and get the most out of what's possible.
Featured Project
02 — Strive
— Strive
A place to stay on your path.
Strive is a social platform for self-improvement. Whether you want to set a personal goal, join a community, or just connect, Strive is built to help you grow and stay motivated. Share your progress, find people working on similar goals, or just meet others working on themselves.
I started building Strive after noticing I’d set big goals and lose momentum over time. Around the same time, I deleted my old social media. What stood out to me — aside from the time it gave back — was realizing how much I had tied my identity to being seen online.
Without it, I noticed how often I’d compare myself to strangers without meaning to. I caught myself doing things partly so I could post about them, instead of being in the moment.
Strive is what I wish I’d had then: the same gravity that makes social media work, pointed at genuine self-improvement and real community.
It’s about making it easier to come back and start again, and connecting with people who are on the same path — no matter where you’re starting from.
— What I’m doing on it
- Founded the company
- Designed and built the mobile app
- Shipped the marketing site
- AI video pipeline driving marketing
- Running the day-to-day
— Stack
React NativeExpoExpo RouterTypeScriptTanStack QueryZustandReanimatedSupabasePostgres + RLSSentryEAS / CINext.jsTailwindVercel
— Where it started
Strive started as a class project I shipped on Android — same idea, simpler shape. I called it CommitPal. The original demo and source are still up if you want to see the first cut.
Where I’ve worked
03 — Experience
Dayton, OH
- Sole backend engineer on multiple mission-critical government applications — owning end-to-end from requirements to production.
- Built REST API endpoints and backend microservices, deployed across local, dev, and production environments.
- Tuned Elasticsearch analyzers, scoring functions, and field mappings to lift search relevance at scale.
- Built a RabbitMQ-based messaging layer between API and worker pods for async, event-driven notifications.
- Wrote a K3D bootstrap that cut local dev environment setup from 60 minutes to under 20.
- Built an interactive launcher to selectively start backend services and a Postgres reset tool to resolve Alembic migration drift.
- Owned production hotfixes, data migrations, and large-scale corrections end-to-end — diagnosing, scripting the fix, and validating in production.
- Lead daily standups to coordinate progress and unblock PRs; technical point of contact for government clients, demoing applications and gathering feedback for the next iteration.
PythonFastAPIElasticsearchRabbitMQPostgreSQLDockerK3DDallas, TX
- Wrote complex SQL queries and data extraction pipelines surfacing workforce insights for stakeholders across multiple business units.
- Built an internal automation service that generates, validates, and distributes offer letters at scale, replacing a manual workflow.
SQLPythonAutomationBoston, MA
- Developed a Java-based inventory tracking system with automated expiry detection, reducing inventory loss and improving supply chain traceability.
- Implemented real-time alerting for low-stock and expiring components, notifying teams proactively and preventing production delays.
JavaInventoryAlertingPalo Alto, CA
- Trained OCR and label-classification models in Keras, hitting 87% accuracy on patient intake forms and cutting manual data entry by ~60%.
- Wired the trained model into a cross-platform mobile app via REST so users could photograph forms and get digitized text in real time.
KerasTensorFlowOCRRESTMobile
Tap a role to expand
Education
04 — Education
UT
ActiveThe University of Texas at Austin
M.S. in Computer Science
Coursework: Deep Learning, Machine Learning, Android Programming.
Jan 2025 — Dec 2026
GPA 3.91
CMU
Carnegie Mellon University
Graduate Certificate — ML & Data Science
Recommender systems, optimized ML pipelines, PyTorch, TensorFlow.
2024 — 2025
GPA 3.84
UTD
The University of Texas at Dallas
B.S. in Computer Science
Academic Excellence Scholar (Full Ride).
2019 — 2022
GPA 3.82
— A few things to look at
Other Projects
05 — Projects
Transformer From Scratch
2025
UT Austin · Deep Learning
Built the full transformer architecture in PyTorch from the ground up — multi-head attention, positional embeddings, residual connections, layer norm. Trained end-to-end and benchmarked across attention configurations.
PyTorchTransformersAttention
ML for Financial Forecasting
2025
UT Austin · Machine Learning
Forecasting stock-price movement on Nvidia and Tesla. Trained and compared decision trees, random forests, and gradient-boosted models against linear-regression baselines, then wrote up methodology and results end-to-end.
Time-seriesMLPython
QA System on SQuAD
2025
CMU · 11-637 FCDS
Closed-domain question answering built up an escalating stack — Jaccard overlap → tf-idf → logistic regression → BERT — comparing accuracy and latency tradeoffs at each step.
NLPBERTPyTorch
CNN Image Classifier (Deployment)
2025
CMU · 11-637 FCDS
Trained and deployed a convolutional network for image classification, comparing CPU vs. GPU training cost and inference latency for the deployment pipeline.
PyTorchDeep LearningDeployment
Airbnb Analytics Site
2024
CMU · 11-637 FCDS
Pulled Airbnb listings and built a responsive helper site that computes analytic results on the fly. Picked data structures and caching strategies based on access patterns at scale.
Data structuresCachingWeb
Collaborative Filtering Recommender
2024
CMU · 11-637 FCDS
Built a collaborative-filtering recommender over product reviews. Compared sparse matrix representations and their storage / runtime tradeoffs.
RecsysNumPySparse
Climate + Food Production EDA
2024
CMU · 11-637 FCDS
Joined climate and food-production datasets to surface correlations and prep features for predictive modeling. Heavy on cleaning, viz, and exploratory work.
EDAPandasViz
OCR Patient Intake Pipeline
2022
Onymos Internship
Trained Keras OCR + label classification models (87% accuracy) to digitize hand-filled patient intake forms; wired through REST into a cross-platform mobile app.
KerasOCRMobile
Get in touch
06 — Contact
Let’s talk.
Quickest way to reach me is email. If it’s about Strive, mention that — I always make time for it.
© 2026 David Denny