Full Stack Developer · Tech Lead · Costa Mesa, CA
Current tech lead on FedEx's First & Last Mile Manifest Platform — Java 17 / Spring Boot microservices on Pivotal Cloud Foundry generating DOT-compliant hazmat manifests and driving an international rollout.
Tech Stack
Backend
Frontend
Infra & Tooling
I'm a Full Stack Developer and Tech Lead at FedEx, where I've grown from FSW I to FSW III since 2021 — building and owning systems that station couriers across the country depend on every day.
My core work is the FedEx FLM (First & Last Mile) Manifest Platform — a suite of Java 17 / Spring Boot microservices on Pivotal Cloud Foundry that station couriers depend on to manage hot and cold space storage retrieval and generate the documents they need to do their job. The platform produces JSON API responses and PDF manifests — Detail, Summary, Runsheet, eManifest, and fully DOT-compliant Hazmat manifests — backed by Oracle, IBM AS400, TIBCO JMS messaging, Dell EMC ECS object storage, and Okta OAuth2 security across all services.
As tech lead I drive architecture decisions, lead code reviews and sprint ceremonies, and coordinate the platform's international rollout as FedEx expands FLM operations globally. I previously managed an overseas development team to help scale the platform during an earlier phase of growth.
Outside work I build community-focused tools — a bilingual youth soccer club platform for Escuela de Futbol GDL-Razo in Costa Mesa, CA, and a live currency rate web app used by a local Santa Ana barbershop and its customers daily. I'm a Swarthmore College Computer Science alumnus and QuestBridge Scholar.
Tech Stack
FedEx First & Last Mile · Costa Mesa, CA
REQUESTED → GENERATED → PRINTQUEUED → ARCHIVEQUEUED → ARCHIVED lifecycle events, consumed by downstream Print, Archive, and Notification microservicesFedEx First & Last Mile · Costa Mesa, CA
FedEx First & Last Mile · Costa Mesa, CA
Ocean Renaissance, LLC · Remote
Swarthmore College Library Lab · Swarthmore, PA
Aerostealth Engineering · Southern California
B.A. Computer Science
2015 — 2019 · Swarthmore, PA · GPA 3.1
Liberal arts college known for rigorous academics. Computer Science curriculum with strong emphasis on theory, systems, and applied machine learning.
Student Athlete & Leader
2015 — 2019
Balanced academics with extensive campus involvement across leadership, athletics, and community.
Recent Work
Community · Web Platform · Bilingual
A full-featured bilingual (EN/ES) website for a youth soccer club in Costa Mesa, CA. Serves 48+ active players across Sub-10 through Sub-16 divisions — with roster management, match scheduling, summer league registration, a photo gallery, and a contact form. Built mobile-first so parents can check schedules from the sideline.
clubgdl.netlify.app →Automation · Finance · Community
Real-time MXN/USD exchange rate web app for a local barbershop in Santa Ana, CA. Customers visit the site — or text/call a dedicated number — to get the day's live rate before exchanging cash. Backend auto-fetches and updates the rate on a schedule, eliminating manual lookups.
claucuts.com →Open Source · Portfolio
This site — a handcrafted portfolio hosted on GitHub Pages / Netlify. Built with intentional typography, CSS scroll animations, and clean markup. No frameworks, no build step.
github.com/ilomeli450 →Academic Work — Swarthmore College
Machine Learning · Computer Vision
Investigation of k-Nearest-Neighbor classifiers vs. single-layer neural networks for handwritten digit classification. Built with PyTorch; explored hyperparameter tuning (hidden layer size, mini-batch gradient descent, Adam optimizer). Single-layer NN outperformed KNN with 98.1% accuracy.
Computer Vision
Explored laplacian pyramid construction for smooth image blending and hybrid image generation. Implemented an interactive face-blend tool using OpenCV that lets users capture their face via webcam and blend it into a composite output.
Computer Vision
Built a system to threshold video frames using HSV color space, apply morphological operators (Gaussian blur + dilation) to remove noise, and track connected components over time. Used to track an orange projectile's trajectory frame-by-frame.
Deep Learning · Graphic Design
Parallel-branch CNN autoencoder architecture to learn content (word images) and style (texture images) representations simultaneously and blend them unsupervised. Novel approach to automated poster generation — trained on 9M word images and 5,500 texture images.
Read Paper →Full Stack · Chrome Extension
Chrome extension journaling dashboard built with React, Redux, and Firebase. Features Google OAuth for private per-user entries and a rich text editor via Draft.js. Inspired by the Momentum browser extension; built with two collaborators.
GitHub →Data Structures · C++
Custom separate-chaining hash table implementation achieving O(1) average time on all core operations. Used to power a Scrabble word assistant that finds anagrams, power sets, and all legal tile plays from a given rack. Included a full performance analysis comparing good vs. bad hashing functions.
Game Development
2.5D side-scrolling game built in Unity using C# and the SCRUM methodology. Environmental message embedded in gameplay — confronting players with the consequences of ocean pollution. Full game loop with animations and progression.
MVC · GUI · C#
Photo editing application using MVC design pattern and a publish/subscribe callback API. Built with GTK# for the GUI. Users can add filters, layer images, scale/position/delete layers, and save composites — all driven by a clean model–view separation.
2017
Swarthmore Intercultural Center
Honors outstanding scholarly work that interrogates social justice, equity, and community principles. Recipient demonstrates engagement of praxis to connect theoretical frameworks and social justice practice.
2016
Swarthmore College
Provides financial assistance for a junior or senior who has demonstrated outstanding leadership qualities at Swarthmore College.
2015
Swarthmore College
Awards eligible students of high academic performance mentorship and intern opportunities to advance their scholastic, inter-personal, and professional success at Swarthmore and beyond.
2015
QuestBridge Foundation
Merit-based national scholarship recipient. QuestBridge connects high-achieving, low-income students with leading universities and full four-year scholarships.
January 25, 2018
On the art of asking good questions as a software engineer — why how you ask matters as much as what you ask. A look at the difference between questions that teach and questions that just get answered, with examples from Stack Overflow and Eric Raymond's framework for smart technical inquiry.
Whether it's a new opportunity, a collaboration, or just a conversation about logistics tech, Java backends, or soccer — my inbox is open.
✉ivanlomeli450@gmail.comEmail ⌥github.com/ilomeli450GitHub inlinkedin.com/in/lomeliiLinkedInCurrently Full Stack Developer III and Tech Lead at FedEx. Open to senior and lead engineering roles — particularly in platform engineering, distributed systems, or high-impact product teams.
January 25, 2018 · Software Engineering
We all know failure is part of success. All the people we revere as "successful" have, at some point, had difficulty doing something. Like all things human, this probably required them to seek the help of others. The life of a software engineer is made up of many questions and much answer-finding. How a software engineer goes about finding answers, however, is very crucial.
Stack Overflow is an interesting questions and answers forum — a beautiful orchestra of collaboration and community. Whenever a software engineer has questions about programming, it's often the first place they land. I've used it myself countless times. And I've noticed that a question's presentation can dramatically affect the quality of the response it receives.
There are questions that are pedantic and not insightful — they make the answerer uninterested and the asker remains unfulfilled. I believe there are ways to ask questions that benefit both parties. Eric Steven Raymond outlines one such approach in "How to Ask Questions the Smart Way," detailing the preparation process before posting, how to pose the question itself, and how to respond efficiently.
A bad question about sorting algorithms, for example, is one that could be answered by a simple Google search. By posing a shallow question, the asker deprives themselves of a deeper learning opportunity and deprives the community of the chance to give an insightful answer. There's no clear ultimate goal stated — just a surface-level request.
A good question, by contrast, carefully lays out the reasons for the confusion. It portrays that confusion as curiosity. It shows the asker's own attempts at solving the problem. It breaks the question into stepping stones. This kind of question invites the community to share the curiosity and work together — and often uncovers a much bigger, more interesting concept than the asker even knew to ask about.
As seen in both examples, bad and good questions exist. It is in the interest of all parties that a question be laid out carefully, researched fully prior to posting, and framed in a way that is interesting to answer. These simple steps create an engaging environment where everyone learns and grows together.