Jeremiah D

Multi-chain AI-asset P 2 P exchange - verifiable compute, API quotas and model-access rights

• Architected the platform from zero - service topology, data model and on-chain/off-chain boundary - while treating deterministic state transitions, testability and failure recovery as first-class design constraints.
• Built a low-latency matching path with a lock-free order book, price-time priority and idempotent fill settlement at 0 ms median match latency; exercised critical behavior through unit, integration, E 2 E, load and failure-recovery testing.
• Designed multi-sig escrow smart contracts with deterministic release conditions and an attestor quorum, plus a reorg-safe indexer and settlement service reconciling $200 M+ notional in locked escrow.
• Built a multi-chain RPC orchestration layer with automatic fail-over and event-subscription middleware that detects failed settlements, duplicate events and rollback conditions.
• Exposed real-time market data and order-room state over Web Socket and gRPC; used Postgre SQL, Redis, Influx DB and Kafka, achieving 99% terminal uptime with production debugging through metrics, logs and tracing.

On-chain data and indexing infrastructure

• Owned a multi-chain RPC scheduler, contract event listeners and an asset reconciliation service in Go, Gin and Postgre SQL, serving real-time indexes to subgraph and webhook customers.
• Tuned a Kafka ingestion pipeline for high-volume chain data and handled chain reorganisations, duplicate delivery, asynchronous ordering and replay idempotently.
• Cut ingestion-to-query latency p 99 by approximately 38% by profiling hot paths, replacing JSON-RPC batchers with a streaming decoder and moving reorg detection off the write path.

NFT orderbook and aggregator infrastructure

• Developed wallet-signature verification, gas estimation and aggregation-routing APIs in Node. js/Type Script; integrated Ethereum, Polygon, Base and Arbitrum with fallback RPC providers and circuit-breakers.
• Wrote integration tests against forked mainnet with Foundry, covering order cancellation, Seaport fills, royalty enforcement and failure paths across smart-contract adapters.

On-chain analytics and quantitative data infrastructure

• Built block crawlers and transaction-parsing services in Python and Go, normalising Ethereum, BSC and Solana data into Snowflake and Click House.
• Added an idempotent backfill orchestrator with checkpointing and dead-letter queues, enabling deterministic recovery and replay after RPC outages.

Fintech payment, ledger and reconciliation systems

• Built payment-ledger and reconciliation services on Node. js, Postgre SQL and Redis with double-entry accounting, idempotency keys and exponential-backoff retries for ACH and wire reliability.
• Improved API response time approximately 30% by tightening indexes and replacing N+1 reads with a batched loader; debugged consistency and retry behavior across service boundaries.

Compensation-data SaaS (YC W 19)

• Started on a React/Type Script dashboard, then moved server-side to own the API gateway and auth service: JWT rotation, RBAC scopes and a webhook-signing helper used across the SDK.
• Worked across browser and API boundaries, translating user flows into reliable authentication, role-permission and session behavior suitable for automated regression coverage.

Financial-modelling SaaS (YC W 20)

• Wrote PRDs, prototypes and roadmaps for a runway-modelling product; translated interviews with 20+ founders into engineering specifications and risk-based acceptance criteria.