Curriculum Vitae

Email verification + domain-success prediction

• Rewrote in-lookup email verification to run in parallel with explicit candidate prioritization.
• Built a full AI-driven domain-success tracking system, doubling prediction success.
• Drove lookup success rate from ~55% (start of 2024) to ~90% (end of 2025) — direct credit/revenue multiplier.

Contact-data normalization

• Solo-shipped migration of 4B+ contact records from nested JSON on the profile model into normalized profile_email and profile_phone tables.
• Real-time sync via low-level Django field-type overrides + custom managers; in-memory dedup; persistence on profile.save().
• Launched with zero downtime, zero performance regression. Unblocks GDPR removals and monolith decomposition.

5x5 data-partner integration

• Highly efficient Airflow DAG built almost entirely on Redshift merge joins; ETL processes 150M record updates per month in 4 hours.
• Pattern (Redshift compute → S3 → Lambda → queue → workers) adopted as the company-wide template for high-throughput data work.
• Privacy-aware pre-filtering by hashed contact comparison in-query — no PII persisted.

AI data-enrichment service

• Built the scalable backend AI enrichment service that opened a new upsellable product class (API enrichments).
• Hackathon prototype that the company subsequently picked up as the multi-quarter GenAI roadmap.

Privacy-removals system

• Built the GDPR-compliant profile-removal system, made possible by the earlier contact-data normalization (contact-to-profile reverse lookup via hashed contact data).

RocketVerify — custom async SMTP

• Designed and led a scalable email-verification service implementing a custom async SMTP protocol; scaled to ~500M emails/month.
• Heavy Playwright integration for the long tail of providers that resist standard SMTP probing.

Lookup parallelization & offline verification

• Reimplemented lookups multi-threaded; reduced P90 latency from ~12s to ~3s.
• Built offline verification on top of RocketVerify with ML-based candidate prioritization; cached the most expensive part of lookups ~90% of the time.

Multi-provider LLM gateway (with Cam)

• Co-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: industry classification, AI enrichment, domain-success prediction, AI-sampled rules extraction, privacy tooling.
• Owns routing, failover, rate limits, retries, cost accounting, secret rotation, and unified observability.

Industry classification — LLM labeling

• Redesigned industry categorization using vectorized semantic industries + LLM-based labeling; lifted fill rate from ~10% to ~99%.

Orphan-profile trimming

• Heuristic + ML scoring methods to trim indexed profiles. Improved search latency and reduced Elasticsearch index size.

Redshift warehouse

• Stood up the company's first data warehouse. Ran a Redshift Serverless pilot directly with the CTO and our AWS rep; killed the pilot for being a poor fit for our workload shape, launched provisioned instead.
• Re-implemented sitemap generation on Redshift with merge-join geometry; 8 days → 4 hours (~48× speedup). Directly load-bearing for SEO, the primary top-of-funnel acquisition channel.

Microservices: SpaCy + Profile-Photo

• Co-designed scalable FastAPI microservices. Authored most of the Terraform.
• Solo-led the Profile-Photo microservice; debugged and resolved memory-leak issues in the facial-recognition stack via Python memory profiler.

Brightdata phone enrichment

• Designed scalable system for Endato data ingestion + matching; produced 32M+ new and boosted phones; substantial Terraform; dashboard for ingestion progress.

Emergency SendGrid replacement

• Replaced SendGrid with a message-queue model (SES + SNS + SQS) processing email activity in real time.

PeoplePro email enrichment

• Overhauled the PeoplePro pipeline; added 14M+ new emails (10M more in flight). Source accounts for ~30% of phones sold.

Academic — entity resolution package

• Built a Python package implementing entity resolution as BERT embeddings + multilayer perceptron. Advised by a JHU CS professor.
• Read dozens of papers on deep-learning ER; benchmarked embedding families and similarity geometries.

Analytics + data quality

• Overhauled finance/revenue ETLs into a real-time customer-events model across all payment processors (Recurly, Stripe, Adyen, Braintree); reduced ETL errors to ~0%.
• Created the data-engineering interview assessment; led 3-month onboarding of three new engineers.
• Migrated all analytics code out of the rr monolith; built custom query wrappers, job scheduling, Slack alerting, and dev tooling.
• Overhauled enterprise attribution; first accurate enterprise revenue, cash, and deal-flow reporting.

Core data

• Sourced, QA'd, and integrated 10+ datasets for UK and French people data — added 2M+ emails and phones.
• Built DataPerson / DataPersonManager abstractions, dramatically cutting time to integrate a new dataset; substantial query-optimization and indexing work.
• Designed a precursor to slowly-changing dimensions via snapshot + diff.

Analytics infrastructure (zero → one)

• Built the company's initial OLAP database / analytics warehouse. Authored the first full suite of product reporting.
• Built attribution infrastructure linking individual transactions and subscriptions across Braintree, Stripe, Recurly, and Adyen back to users; first cash-to-user attribution at the company.
• Optimized + scaled ETLs to handle a significant volume increase, ~50× ETL performance improvement.

Data-quality + data-partner work

• Sourced hundreds of contact-data providers; helped negotiate and close new provider deals.
• Drove integration of the PeoplePro 2021 dataset (~16% of all phones in the database).

Lead Junior Analyst — Resolve Growth Partners

Founder & President — Johns Hopkins Quantitative Finance Society

Woodrow Wilson Research Fellow — Johns Hopkins University

Partner / Associate — A-Level Capital

Data Analyst — Massachusetts Land Company

Data Science Intern — Empower Schools

Data Science & Bioinformatics Intern — Curoverse Inc. (acq. by Veritas Genetics)

Executive Delegate (Internship) — U.S. Department of Education — Massachusetts State Student Advisory Council