Open to software, product, and applied AI engineering roles · New Jersey, USA

Jefrey Peralta · Software Engineer · Full-Stack Systems · Applied AI

I turn operational problems into production software and AI systems.

I spent years operating businesses that ran on spreadsheets, CRMs, and duct tape. Now I build the software those operations deserve: payment infrastructure, internal tools, and AI systems, carried from discovery and architecture through deployment and operation.

View deployments ↓ResumeView GitHub ↗

Production fintech platform, launched and operated live

Payments, payouts, and reconciliation shipped

Internal operations software in daily use

AI features running in a real product

Deployments

Systems built inside the operations they run.

Velocity admin Command Center showing revenue, trader accounts, pending payouts, risk flags, platform health, and daily operating priorities, populated with synthetic data
The admin Command Center, one operating view of the whole business. Synthetic data.

Production deployment · Financial operations

Velocity Platform

Operated in production, May to July 2026

The problem
Running the evaluation business required checkout, trader accounts, market data, risk enforcement, payments, payouts, and internal operations to work as one system.
The outcome
One production platform handled the customer and operational lifecycle from signup through trading and payout review.

RoleFounder, system architect, and sole engineer

Next.js · TypeScript · PostgreSQL · Supabase

Read the case study
ResolveOS opportunity pipeline organizing settlement candidates by stage with amounts, funders, confidence weights, and next actions, populated with synthetic data
The settlement pipeline: offers tracked by stage before they become payment plans. Synthetic data.

Embedded deployment · Commercial collections

ResolveOS

Used daily in a live collections workflow

The problem
Collections work depended on spreadsheets, memory, unwritten follow-up rules, and manual payment tracking.
The outcome
ResolveOS turned that workflow into a daily operating system for priorities, settlements, payment schedules, forecasting, and reconciliation.

RoleEmbedded engineer, operator, and primary user

Next.js · React · Supabase · Tailwind CSS

Read the case study

Additional engineering work

Supporting work · Marketing and acquisition

Live public website

velocityfunds.io

The public face of Velocity Funds: a scroll-driven launch site that is also real infrastructure, with lead capture wired into the platform CRM and consent-gated analytics. Live, and you can click it.

  • Five-act scroll-cinematic homepage with GSAP ScrollTrigger and lazy-loaded three.js
  • Server-side lead capture proxying into the platform CRM with a service token
  • First-party consent framework gating Meta/Google/X pixels
Next.jsTypeScriptTailwind CSSGSAPthree.jsVercel

Case study →Live site ↗

Operating model

How I deploy.

Every system in this portfolio went through the same six stages.

  1. 01

    Discover

    Understand users, workflows, constraints, the technical environment, and the business outcome that matters.

  2. 02

    Scope

    Reduce ambiguous problems into concrete requirements and measurable success criteria.

  3. 03

    Architect

    Design the data model, APIs, integrations, AI layer, infrastructure, security, and rollout.

  4. 04

    Build

    Implement across frontend, backend, data, integrations, infrastructure, and AI.

  5. 05

    Deploy

    Integrate into the real operating environment, observe usage, handle failures, iterate.

  6. 06

    Measure & generalize

    Evaluate adoption and impact; turn what worked into reusable components and systems.

Applied AI

Applied AI, not AI demos.

The AI work I care about runs inside real operational systems, where reliability, permissions, data quality, cost, and latency all have consequences.

Trader analytics in production

VI Pulse, Sentinel, and Debrief score trading performance, watch risk rules, and review sessions over real account data in the Velocity platform.

OpenAI API · 13 versioned prompts · model router · SSE streaming

Behavioral intelligence for operators

Retention Intelligence turns trader behavior into health scores and a ranked worklist, so an operator knows who needs attention today.

behavioral signals · health scoring · admin worklists

AI-assisted engineering

I build with coding agents inside a disciplined loop: repository context, scoped plans, line-by-line diff review, and verification in the running app.

OpenAI Codex · Claude Code · context engineering · human-in-the-loop review

Evaluation harnesses, retrieval over operational data, and agent workflows are next on my list. They'll appear here as case studies once they ship.

Experience

Operator first. Engineer because of it.

  1. 2025 – Present

    New Jersey (remote)

    Velocity Funds / Velocity Co.

    Founding Product Engineer

    Designed, built, launched, and operated the platform behind a futures evaluation business: payments, trading, risk controls, payouts, internal operations.

  2. 2024 – Present

    Advanced Recovery Group

    Account Manager

    Embedded in a live collections operation: I perform the workflow, and built ResolveOS to replace the spreadsheets and memory it ran on.

  3. 2021 – 2024

    Utopian Estate / Nova Resolutions Group

    Founder & Managing Partner

    Ran service businesses on CRM systems and automations I designed and built.

  4. 2019 – 2021

    Newark, NJ

    Leverage Companies (Private Equity)

    Lead Acquisition Manager & CRM Systems Developer

    Designed Salesforce architecture and workflow automation, translating leadership's business requirements into CRM systems that scaled operations.

Education

Western Governors University: B.S. and M.S. Computer Science pathway, currently enrolled, January 2026 – Present.

Berkeley College: B.B.A. · Salesforce Certified Administrator

Capabilities

What I can build and own.

Operational discovery and scoping

Requirements drawn from doing the work: I've run the collections, trading, and CRM workflows I build for.

collections ops · trading ops · CRM ops

System architecture and integration boundaries

Data models, service boundaries, queues, webhooks, and third-party integrations designed as one topology.

PostgreSQL · BullMQ · webhooks

Full-stack application engineering

Dashboards, admin surfaces, and typed APIs that people work in all day.

React · Next.js · Node.js · TypeScript

Payments and financial workflows

Checkout, webhooks, entitlements, payouts, and reconciliation built to survive retries.

Authorize.Net · PostgreSQL · Redis

Data modeling and PostgreSQL systems

Schemas written as migrations, row-level security, money in integer cents.

PostgreSQL · Supabase · SQL

Applied AI systems

LLM features shipped inside real products: versioned prompts, model routing, usage caps, and streaming.

OpenAI API · SSE streaming · versioned prompts · model routing

Business systems and integrations

Salesforce and CRM architecture, email infrastructure, market data, and the automation between them.

Salesforce · Resend · protobuf over WSS

Deployment, operations, and reliability

Deploys, monitoring, incident response, and the runbooks that follow.

Vercel · Railway · Docker · Sentry

Need an engineer who can own the messy parts?

I'm looking for software, full-stack, and product engineering roles where I can build reliable systems and keep growing, including applied AI and forward-deployed work. I bring the domain context, architect the system, build it, deploy it, and take responsibility for what ships.

Start a conversationLinkedIn ↗GitHub ↗Resume