If 2023 was the year of AI demos / POCs, 2024 is the year of reliable AI Voice Agents.We help engineering teams make their AI Voice Agents reliable. We are hiring!

I consult and advise startups + public companies on data science and growth strategy

I ran a growth-y data science team @ Citizen. We supported the following functions: product, marketing, expansion, policy, and operations. In addition to supporting other functions, my team directly ran 5+ weekly growth and revenue-increasing experiments (pricing tests, onboarding experiments, feature launches, etc.)I 2xed the team size to 6 directs. My team helped set many OKRs for the entire company but directly owned several critical top-line growth and revenue OKRs.Few wins from my team- 3xed CTR, improved user retention by 5% abs- Ran 20+ onboarding experiments that yielded $2M+ in incremental ARR- Saved 350k / year in infra costs- Kick-started ML personalization efforts to improve user retention (est. impact: 5-10% increase in retention)

I joined Citizen as its 2nd DS. I helped 4X MAU, drove strategic discussions with board members around pricing + fundraising, and sharpened the team's product/growth thinking through a metrics-first lens.Highlights- Made capital allocation decisions to fund or de-fund entire teams by comparing growth vs. revenue trade-offs- Helped set OKRs for the entire company for multiple quarters- Created a metrics architecture that allowed each person in the company to connect their work to concrete business impact (this drove employee engagement and ownership)- Pushed for COVID coverage in March of 2020 within the product; increased DAU by 20% and lead to millions of users being more safe and informed- Supported fundraising efforts by crafting our growth story from metrics pov (acquisition, activation, retention, virality)

Built machine learning-based systems that drove millions in revenue, increased gross margins, and optimized efficiency by 10x across the org. Worked with the C-Suite to lead high-performance teams and execute critical, high-profile business initiatives.Notable wins:1) Saved $2M / year and drove⭡$XXXM / year in additional revenue (without an increase in sales spend)2) Simplified supply chain so the factory can build +10% more cars + build the right cars that people want by forecasting demand (⭡Working capital + Gross margins)3) Improved delivery experience by giving owners a better ETA of when their cars will be delivered (Better delivery experience ⭢ More referrals ⭢ More revenue with lower CAC)

Machine learning @ Growth; see⭡for the wins

Content based techniques for medical image retrieval. Techniques used: stacked autoencoders (deep learning), LSH / k-NN, various classifiers (SVM, random forests, etc.) and domain heuristics.Lead author in published paper: https://arxiv.org/pdf/1610.00320.pdf

Domain modeling, API design, reactive backend services (user management & auth, time series, biz logic), devOps (scala, docker, kafka, akka, elasticsearch, mongo)Co-­developed datacenter simulator (akka). Benefits: more robust, exhaustive & cheaper testing; also what­if scenariosCo-­developed power optimization algorithms (core IP) which can reduce the total cost of ownership of datacenters by up to 50%; developed L1 regression models to calibrate current/voltage sensors

Lead Reflektion’s search and relevance team of four engineers (reported directly to Chief Data Scientist); deployed new products to customers such as Converse.comDesigned and implemented Reflektion’s search product (API, backend services); increased customer engagement by 70% and increased conversion for our customers by 26% (python, pyramid, elasticsearch, pig)Developed new ranking techniques combining pCTR (logistic regression)and domain-specific heuristicsDesigned a multivariate testing framework using a multi-armed bandit approach to optimize the CTR and overall engagement across an arbitrary number of parameters

Shipped core application features using Rails; A/B testing; featured in CNN & MSN Money.Marketing strategy and customer development with co-founders (Bo & Jon)