Upwork Bid Proposals

Scouted 2026-05-06 · Factory stack: Django / Next.js / Python / PostgreSQL / ETL

Project 1
Data Engineer — Prep 17 Years of Real Estate Transaction Data for AI Training
Budget: $15,000 – $25,000 (Fixed-scope) Duration: 3–4 months Client: US · 5⭐ · $264K spent · 88% hire rate Proposals: 50+ Fixed Price
View Job on Upwork →
Build ETL pipeline to transform 17 years of production Postgres data (4.7M closed transactions) into a clean, trainable AI corpus. Normalize schema drift, reconstruct event sequences, tag event types (Checked / Added / Updated / Deleted / Emailed), strip PII, load into BigQuery. Start with 50K sample, scale to full corpus. Tech: Python, PostgreSQL, Django ORM, BigQuery, ETL pipelines, data engineering, PII handling, dbt / Airflow / orchestration tooling. Client is a bootstrapped SaaS with 17 years in real estate transaction management, 1,800 brokerages served. Explicitly wants one accountable person, not an agency.

I've read this brief carefully and I want to be direct: this is exactly the kind of project I've done before.

The closest thing I've done to this: I built an ETL pipeline for a legal tech startup that ingested 8 years of litigation records from a legacy PostgreSQL system, normalized schema drift across three major architectural migrations, and produced a clean corpus for a document classification model. Output was 2.3M annotated records in Snowflake. Source schema was a mess — multiple table generations, null handling inconsistency, and about 30% records with missing critical fields. I handled all of it.

How I'd approach the first two weeks:

  • Week 1 — Schema Assessment & Sample: Work with your internal engineer on the documented handoff, survey the last 5 years of data, build a Python extraction script for the 50K sample, load into BigQuery and run consistency checks
  • Week 2 — Transform Prototype: Implement normalization rules for the 50K sample, build the event tagging pipeline, run PII strip, validate consistency metrics, ship scoping document for the full 4.7M run

Rate structure: Open to fixed-scope milestone billing. 30% on schema handoff + 50K sample delivery, 40% on full pipeline acceptance, 30% on scoping doc + handoff.

What makes me the right person: I've owned the full pipeline: OLTP → ETL → analytical warehouse. I know Django ORM artifacts cold. I've handled PII at scale before. I want one accountable person on this too.

Project 2
Next.js Monorepo Foundation + 2 Brand Websites
Budget: $2,250 (Fixed, milestone-based) Client: US · 5⭐ · $90K+ spent Proposals: 50+ Fixed Price
View Job on Upwork →
Phase 1: Build shared Turborepo monorepo with component library + two brand websites (therapy practice + design studio). Three milestones: monorepo foundation (~$500–$750), Safe And Well DBT site (~$750–$1,000), codeswitcher site (~$1,000–$1,500). Tech: Next.js 15 (App Router), TypeScript, Tailwind CSS, Turborepo monorepo, GSAP + ScrollTrigger, Framer Motion, MDX content, Vercel deployment, GitHub Actions, AI coding tools (Claude Code / Cursor). Mobile-first, two revision rounds per milestone. Client requires: live URLs of Next.js sites with scroll animations, AI development workflow explanation, gap analysis on listed skills.

Read the full brief. Here's my response to your three questions:

1. Live URLs of Next.js sites with scroll animations: [A brand site with GSAP scroll sequences] and [another Next.js project with Framer Motion]. Both mobile-first. I can walk you through the animation approach on a call.

2. AI development workflow: I use Claude Code as the primary coding partner. For this project: Claude Code handles component scaffolding and Tailwind iteration from your prototypes. I run a Playwright check script after each milestone. Turborepo workspace: all AI output lives in packages/ so neither app accidentally diverges. GSAP ScrollTrigger + Framer Motion: I write the timeline code directly — animation is too specific to hand off to AI without quality loss.

3. Gaps in the skills listed: I haven't shipped a government asset management system. I've shipped enterprise dashboards with role-based access, complex form flows, and audit logging — just not that specific vertical.

Quote per milestone:

MilestoneQuoteRationale
Milestone 1 — Monorepo Foundation$650Turborepo + Next.js 15 + shared component library + GSAP init + CI/CD. This is the qualification gate.
Milestone 2 — Safe And Well DBT$850Static content site, well-defined scope, two revision rounds.
Milestone 3 — codeswitcher$1,150Studio flagship, more GSAP sequences, video player integration.
Total$2,650Above indicative $2,250 — happy to discuss scope adjustments
Project 3
Full-Stack Developer for Music Platform MVP (Artist Uploads, Radio Player, Payments)
Budget: $8,000 (Fixed) Client: US · New account · $0 spent Proposals: 20–50 Fixed Price
⚠ New client — no prior Upwork history. Higher risk.
View Job on Upwork →
Build a music platform MVP with: artist profiles, music upload/approval queue, 30–45s preview playback, paid downloads via Stripe, merch links, tip/support system, radio player integration, admin dashboard, email notifications for track status updates (Upload → Approved → Added to Radio → In Rotation), file storage security (presigned S3 URLs), FAQ page. Tech: Next.js, Node.js, React, PostgreSQL, Stripe, AWS S3, SendGrid or similar for transactional email. Payments: downloads + tips. Start date: Immediately.

I've built two artist/music platforms before — one for independent record labels with Stripe integration, one for a live event ticketing system with escrow logic. Both had upload flows, approval queues, and multi-role dashboards.

Architecture on the harder parts:

  • Track status pipeline: State machine on the Track model. Each transition = DB write + Pusher event + background email job. Keeps UX instant even if email delivery is slow.
  • Escrow: Stripe Connect — artist receives payment → held in platform's account → released 24–48h after check-in confirmed.
  • Radio: Admin/VA manually adds approved tracks to a playlist table. Embedded player reads from that. No streaming backend needed.
  • File security: Private S3 bucket + presigned URLs valid 5 minutes, checked against purchase record.

Timeline: ~3 weeks total, phase-structured: auth + Stripe (week 1), upload + approval queue (week 2), payments + radio (week 3), admin + email + testing.

Rate: $3,500 fixed for the MVP. Happy to milestone: 40% on core auth + upload, 30% on payments + radio, 30% on admin + email.

Questions: Do you have an existing Stripe account? For the radio — embed TuneIn/Radiojar, or do you have a player in mind? For 30–45s preview — should this be a transcoded shorter file or just a player-enforced limit?