I build software from idea to production.
I'm Melvin Berkoh, a software engineer focused on full-stack products and frontend systems. I also work across backend logic, databases, testing, and deployment.
Engineering focus
The areas I spend most of my time working in.
Full-stack products
Frontend, APIs, auth, and application logic.
Frontend systems
React, TypeScript, responsive UI, and accessibility.
Backend & data
Node.js, PostgreSQL, Prisma, Python, and SQL.
Product ownership
Build, debug, test, document, and ship.
How I work
I like owning features end to end. I want to understand what is happening across the UI, backend, database, and deployment.
Selected builds
Products I took past the idea stage.
These projects go past the UI. I worked on the backend, database and deployment too.
OpsDesk
Support and incident management built around real multi-user workflows.
OpsDesk handles services, tickets and incidents inside multi-tenant workspaces. Permissions are enforced on the server. SLA deadlines are stored with the work they belong to.
Tenant boundaries
Workspace-owned data is scoped on the server.
Authorization
Authentication identifies the user. OpsDesk decides what they can do.
Workflow logic
Tickets can become incidents with stored SLA deadlines.
Application Tracker
A full-stack app for keeping a job search in one place.
Applications connect to interviews and follow-ups. Activity history keeps track of what changed. Job post URLs can also be imported before an application is saved.
User-owned records
Application data stays scoped to the signed-in user.
Hiring workflow
Applications connect to interviews and follow-ups.
Job post import
URLs are checked before remote content is fetched and parsed.
Technical work
Different problems need different tools.
These projects pushed me outside a normal full-stack app. PolicyScope started inside the browser.
PolicyScope
A Chrome extension that finds important policy clauses before you agree to them.
PolicyScope scans Terms of Service and Privacy Policy pages. It groups detected clauses by category and highlights them on the page. Users can also request a plain-English explanation.
Clause detection
Rule-based detection keeps the main scan fast and predictable.
Secure AI flow
AI requests go through a Node and Express backend. The API key stays out of the extension.
Plain-English explanations
AI is used on demand. It is not required for the main detection pipeline.
Privacy Policy
Example Service
We collect information that you provide when creating an account and using the service.
We may share personal information with service providers and affiliated partners when needed to operate the platform.
Your subscription renews automatically unless cancelled before the next billing date.
Any legal dispute will be handled through binding arbitration and limits on class actions.
We may update these terms from time to time. Continued use of the service means you accept the updated policy.
commerce-pulse.pipeline
Olist analytics workflow
Data flow
CSV
Raw Olist data
Python
Profile + ingest
PostgreSQL
Warehouse
dbt
Transform + test
SQL
Business analysis
Tableau
Dashboards
Reporting layer
Historical Olist dataset
$13.22M
Product revenue
96,478
Delivered orders
4.16
Avg. review score
8.11%
Late delivery rate
Delivery vs. review score
One finding from the reporting model
Warehouse layers
CommercePulse
Raw e-commerce data turned into a reporting warehouse and business dashboards.
CommercePulse starts with raw Olist CSV files. Python profiles the data before it moves into PostgreSQL. dbt handles transformations and tests. SQL and Tableau sit on top of the final reporting models.
Layered warehouse
Raw data moves through staging, intermediate models and reporting marts.
Data quality
Payments, reviews and geolocation data are cleaned before reporting.
Business reporting
Final models support revenue, delivery, customer and seller analysis.
Dashboard metrics shown here come from the historical Olist dataset used in the project.
Journey
From first opportunity to trusted ownership.
Each step gave me more responsibility. I went from building pages to owning product flows, system logic and data.
Summer 2023
First industry experience
Mendham Township
Worked as a Website & Forms Intern on real municipal systems. I redesigned CivicPlus pages, built forms, worked with analytics and documented workflows for staff.
Summer 2025
Frontend internship
Eligido
Joined Eligido as a frontend intern and built its public-facing React landing page. I turned complex startup material into a responsive site that was easier for users and stakeholders to understand.
May 2026
Education + capstone
Graduated from NJIT
Finished my B.S. in Computer Science. PolicyScope placed 3rd overall in the senior capstone showcase, where I focused on clause detection and backend functionality.
August 2026
Return engagementReturned as a contractor
Back at Eligido
After my internship, Eligido brought me back as a paid Frontend & Trust Layer Engineer. My work expanded into frontend architecture, interaction flows, accessibility and the Stage 1 product experience.
August 2026
Independent product build
Application Tracker
Built a multi-user job search app with authentication, relational data, interviews, follow-ups and job-post imports. It became a product I could use during my own job search.
Aug – Sep 2026
Systems
OpsDesk
Pushed deeper into backend architecture with multi-tenant workspaces, server-side RBAC, incident escalation and SLA logic. System boundaries became a bigger part of how I designed software.
September 2026
Data
CommercePulse
Built an analytics pipeline from raw CSV files to PostgreSQL, dbt models, SQL analysis and Tableau dashboards. It brought the same end-to-end mindset into data engineering.
NOW
Still building.
Technical stack
Built with tools I actually use.
Explore the stack to see where each skill shows up in real work.
01 / Interface
Frontend
Interfaces that stay clear, responsive and usable as the product grows.
Explore the stack
02 / Application logic
Backend
Server-side rules, permissions and APIs that keep product behavior predictable.
Explore the stack
03 / Storage + analytics
Data
Relational product data, analytics models and reporting pipelines.
Explore the stack
04 / Shipping
Engineering
The work around the feature: testing, debugging, validation and deployment.
Explore the stack
More work
Different problems.
Different builds.
Other projects that pushed me into client work, multiplayer systems, and data science.
About
I like understanding how the whole thing works.
I started with websites. Over time I became more interested in everything happening behind them.
Now I like working across the interface, server, database, testing, and deployment. I do not need to own every layer on every project, but I want to understand how they connect.
That mindset is what pushed me from frontend work into full-stack products, backend systems, and data.
How I work
Ownership
I like understanding the feature end to end. That means knowing what the UI does, what happens on the server, and what data changes underneath it.
Clarity
I would rather build something another engineer can understand than make a simple problem look complicated.
Real use
The projects I enjoy most solve an actual problem. Real workflows, real constraints, and a reason for the software to exist.
Outside the projects
Education
B.S. Computer Science
New Jersey Institute of Technology
Campus
NJIT Men’s Lacrosse
Worked around a team environment while in school
Professional
Client + startup work
Freelance projects and work with Eligido
The goal
Keep getting better at taking a problem, understanding the system around it, and shipping something useful.
Contact
Got something
worth building?
I'm interested in software work where I can solve real problems, own meaningful parts of the product, and keep growing as an engineer.
Email meDirect line
melvinberkoh@gmail.com