
A modular DeFi product that brought live market intelligence, liquidity discovery, governance, and community mechanics into one readable experience.
Product surfaces
7 modules
Dashboard, market intelligence, analytics, pools, games, lottery, and governance.
Live data breadth
11 tracked tokens
Market feeds are aggregated into a single decision surface with trend signals.
Community mechanics
3 game loops
Play-to-earn concepts were integrated as part of the broader ecosystem.
Trust signals
6 prize tiers
Lottery mechanics and round history were structured for transparent fairness.
The Problem
DeFi experiences often split analytics, pool discovery, governance, and community mechanics across disconnected tools. Nextflow was designed to turn that sprawl into a single product with a clear operational rhythm.
Nextflow started from a simple product problem: most DeFi users bounce between dashboards, explorers, pool trackers, and governance tools just to understand what is happening. The goal was to collapse that fragmented experience into a single product that felt fast, expressive, and coherent.
The platform was built as a modular React and TypeScript application where each domain owns its own surface, data layer, and interaction model. TanStack Query handled cache discipline and background refresh, while the wallet and contract layer abstracted on-chain reads and writes into a predictable developer workflow.
The result is a product that feels market-driven rather than page-driven. Users can scan the system state, inspect performance, discover yield opportunities, and participate in governance without losing context or being overwhelmed by raw protocol complexity.
What changed
Instead of forcing users to stitch together third-party dashboards and protocol tools, the platform gave them one coherent place to monitor markets, compare opportunities, take action, and stay engaged with the ecosystem.
Why it was hard
The challenge was not just data integration. It was deciding how much motion, context, and abstraction to add before the product stopped feeling trustworthy. The system had to feel alive without sacrificing scanning speed or decision clarity.
Constraints
My role
Proof
Real screens from the product — each one supports a specific argument about clarity, control, or observability.
Signal SurfaceClick to view full sizeA landing surface that lets users understand the system state in seconds instead of navigating multiple protocol views.
The KPI and sparkline layout was designed for quick scanning before deeper interaction.
Decision SupportClick to view full sizeRanked tables and charting views translate noisy pool data into comparative product signals.
This surface reduces raw protocol noise into a reading order users can actually act on.
Yield LensClick to view full sizeAPR-oriented pool views make opportunity discovery feel like a product feature, not a spreadsheet exercise.
Derived metrics are surfaced with explicit assumptions so the interface stays actionable and honest.
Community LoopClick to view full sizeGamified modules extend product engagement without feeling bolted on or disconnected from the platform identity.
Motion and card choreography were used to create energy while preserving legibility.
Trust LayerClick to view full sizeProposal visibility, vote progress, and tagging were designed to make governance participation feel understandable.
The goal was to make governance navigable for users who are not already protocol power users.
Decisions
The strongest work is visible in the choices made under pressure, not just in the final interface.
Challenge
The product had too many distinct jobs to force everything into a single interface model.
Decision
Each domain became its own module with dedicated components, hooks, and data contracts.
Tradeoff
This increased coordination around shared patterns, but it prevented the product from collapsing into a brittle monolith.
Challenge
Yield hunting becomes unusable when users have to interpret every pool from first principles.
Decision
APR derivations, high-yield filtering, and comparative analytics were elevated into first-class product features.
Tradeoff
Derived metrics require stronger explanation and validation, but they make the product meaningfully more useful.
Challenge
A Web3 brand can become unreadable if every surface behaves like a launch animation.
Decision
Motion was concentrated around transitions, card focus, and discovery moments while charts and KPI surfaces stayed disciplined.
Tradeoff
More visual tuning was required, but the result feels energetic without losing trust.
Architecture
CoinGecko, DexScreener, and on-chain reads provide the raw market and protocol signals.
→ Fresh multi-source inputs without forcing users into source-level complexity.
TanStack Query handles polling, stale-time strategy, cache reuse, and background refresh.
→ Real-time feel with predictable network behavior and cleaner UI state management.
Markets, pools, games, and governance each operate as focused product modules.
→ Faster iteration and clearer product boundaries across the experience.
wagmi, viem, and ethers abstract reads and writes into reusable interaction patterns.
→ Safer, more consistent user actions across the DeFi surface area.
Dashboards, analytics, and governance views turn protocol complexity into readable UI flows.
→ Users can decide and act without bouncing between disconnected tools.
Nextflow uses a feature-based frontend architecture where each domain owns its own interaction model, data hooks, and UI surfaces. TanStack Query is the operational backbone for live data, while wallet and contract interactions are abstracted into repeatable patterns that keep market exploration and user action in sync.
Product surfaces

The main command surface aggregates live market movement, watchlist signals, and high-level KPIs into one fast-scanning view.
Tech Stack
Framework
UI
Web3
Data
Languages
TypeScriptSolidity (smart contracts)CSS/TailwindWhat this project proves