On building this
A short note on why I built a website now, after five years of not having one.
Johnny Saldana
Backend & ML systems engineer. Founding-era engineer at RocketReach — contact-data infrastructure, email verification at scale, and the multi-provider LLM gateway behind every AI feature the product ships. Currently in Manhattan; relocating to Stanford in September.
Recent writing
Index →Selected work
All →Multi-Provider LLM Gateway
RocketReach
Co-designed and built the in-house service that abstracts over Anthropic, OpenAI, xAI, and Gemini behind a single interface, with native websearch. Backbone for every AI feature at the company.
Providers ANTHROPIC · OPENAI · XAI · GEMINI
Surface ALL AI FEATURES
Concerns ROUTING · FAILOVER · COST · OBSERVABILITY
Contact-Data Normalization
RocketReach
Solo migration of 4B+ contact records from nested JSON on a monolith model into normalized relational tables, with real-time sync, zero downtime, and zero performance regression.
Scale 4B+ ROWS
Downtime ZERO
Mode REAL-TIME DUAL-WRITE
RocketVerify — Custom Async SMTP
RocketReach
Designed and built a scalable email-verification service implementing a custom async SMTP protocol, with Playwright fallback for the hardest cases. Throughput at ~500M emails/month.
Throughput ~500M / MONTH
Protocol CUSTOM ASYNC SMTP
Fallback PLAYWRIGHT
Redshift Warehouse — Pilot, Kill, Re-Launch
RocketReach
Stood up the company's first data warehouse. Ran a Redshift Serverless pilot directly with the CTO and our AWS TAM, killed it for being a poor fit, launched provisioned instead. Sitemap generation went from 8 days to 4 hours.
Sitemap perf 8 DAYS → 4 HOURS · 48×
Surface FIRST COMPANY DW
Workload shape BATCH-HEAVY · NOT BURSTY
Research
All →Deep-Learning Entity Resolution — BERT Embeddings + MLP
Johns Hopkins University
Python package implementing entity resolution as a two-stage pipeline — BERT embeddings of structured and unstructured tuples, followed by a multilayer perceptron classifier. Built with a JHU professor advisor across 2023.
JHU Quantitative Finance Society — Founder & President
Johns Hopkins University · Graduate School of Applied Math & Statistics
Co-founded and led JHU's graduate quant-finance research society as a long-short systematic-trading group with $30K AUM. Built the backtesting engine and led 20+ student researchers including 8 PhD students. WorldQuant Challenge Semi-Finalist.
Implied–Realized Volatility Gap — Woodrow Wilson Research Fellowship
Johns Hopkins University · Woodrow Wilson Fellowship
$10K research fellowship at Johns Hopkins under Prof. Jonathan H. Wright, investigating the persistent gap between options-implied volatility and subsequently realized volatility in equity markets.
Currently
Now →Reading Deep Learning (Goodfellow), going deep on transformer internals, distributed-training primitives, and inference-serving stacks. Shipping a series of public technical write-ups on the backend infrastructure I've built at RocketReach. Open to conversations about ML/systems hybrid roles in the Bay Area starting fall — full reading list and learning targets on Research → In progress.