LEAD / SENIOR PRODUCT ANALYST · ANALYTICS ENGINEER

Full-cycle analytics —
from raw data to the decision.

10+ years, B2C & B2BAdTech · EdTech · StreamingSt. Petersburg — remote

Lead / Senior Product Analyst with 10+ years in B2C and B2B products (AdTech, EdTech, Digital Creative Products). Combines an engineering background, classical statistics, ML, and AI to deliver step-change growth in business metrics and surface product and marketing growth levers.

CONTACT
a.oberon@gmail.com
LinkedIn

Experience

VK · LinguaLeo · Tutu.ru · BI Consult

Ten years of product analytics — from the data platform to the decisions made on top of it.

Senior Product Analyst

Dec 2024 — Present
Video Streaming Service · SVOD (under NDA)
  • Built the analytics function and data platform from scratch. Embedded analytics and the HADI cycle into prioritization and decision-making.
  • Established a HADI + RICE framework for evaluating product initiatives: scored ~250 hypotheses, shipped several Gradio dashboards for what-if ROI modeling.
  • Designed metric systems for funnels across multiple projects: CR, Retention, LTV, time spent, cohort metrics.
  • Identified and validated growth levers for marketing and product: a revamped recommendations block doubled viewing depth and lifted subscription conversion by 9%; automated YouTube channel discovery sourced ~1,800 partner channels, generating +12% leads.
  • Ran ~20 product A/B tests; analyzed 100+ ad channels and campaigns (Google Ads, Meta Ads, YouTube, affiliate partners, organic traffic).
  • Built a PoC of a self-service AI analytics tool; ensured data trust with a custom semantic layer.
A/B Testing · MDE · Unit Economics · ARPPU · Subscription Metrics · Paywall Optimization · Pricing & Packaging · Marketing Attribution · Personalization · Product & Marketing Metrics · Text-to-SQL · Semantic Layer · LLM Pipelines · RAG · Prompt Engineering · AI Agents · Claude / OpenAI APIs · MCP · Vector Search · ANOVA · Python · Jupyter · GA4 · Metabase · dbt · MetricFlow · ClickHouse · BigQuery

Fractional Product Analyst

May 2021 — Dec 2024
Independent · multiple B2C products
  • Advised on BI implementations, data warehouses, data collection, and data-driven decision making.
  • Established KPI systems and uncovered metric relationships, with a focus on unit economics for B2C products. Optimized LTV, retention, and user funnels.
  • Deployed pilot analytics stacks using ClickHouse and open-source BI tools (Metabase, Redash, Superset). Automated data collection and transformation with Python + cron or dbt/Airflow.
  • Built comprehensive data views and dashboards.
  • Proposed growth hypotheses and testing methodologies after analyzing business and product needs. Organized key analytics processes: data quality control, business metric monitoring, incident management, ad-hoc reporting, A/B testing, Scrum, analyst mentoring, and integrating analytics into decision-making and release cycles.
A/B Test Design · GrowthBook · Switchback Testing · KPI Frameworks · North Star Metric · AARRR · Growth Loops · LTV:CAC · Unit Economics · Funnel Optimization · Retention · RFM Segmentation · E-commerce / SaaS / Fintech Metrics · Embeddings · Vector Search · Reporting Automation · Mentoring · Python (pandas, NumPy) · PostHog · AppsFlyer · Firebase · Metabase · Superset · ClickHouse

Product Analyst

Oct 2019 — Apr 2021
VK · Advertising Technologies
  • Ad-hoc analytics for the Ads Technologies and Superapp divisions: optimized the ads manager UI and ad auction algorithm; quantified the impact of ML model quality metrics on business KPIs.
  • Supported product teams by building KPI systems, creating and maintaining Superset dashboards, and analyzing A/B test results. Designed data pipelines using Spark and Airflow to process and load data into ClickHouse, while analyzing various data sources (e.g., MySQL) and tracing business logic in legacy PHP systems.
  • Enhanced analyst productivity: a unified code repository, an event debugging tool, and Python libraries for auto-generating Superset dashboards and bootstrap-based A/B test analysis.
  • Prototyped ML solutions for entity similarity search and community classification using embeddings (ALS, LDA, word2vec) and nmslib.
AdTech · Product Analytics · A/B Testing · KPI Systems · Ad Auction Optimization · Recommender Evaluation · Churn Prediction · Propensity Scoring · ML Model Evaluation · Embeddings (word2vec, ALS, LDA) · Bootstrap Methods · Superset · Python · Spark · Airflow · ClickHouse

Data Analyst / Analytics Engineer

Oct 2018 — Oct 2019
Tutu.ru · Travel
  • Refactored the BI stack and piloted a Metabase/Exasol/ClickHouse integration. Built an ETL engine using Exasol's embedded Lua runtime, and developed a data quality monitoring and anomaly detection system. Gathered business requirements to create dashboards in Metabase and managed ad-hoc data collection from MySQL, ElasticSearch, and Prometheus. Trained analysts on Exasol, Metabase, and ClickHouse.
  • 2× faster analytics workflow · ~300 raw tables in the ETL pipeline · 20 data marts across 5 business domains · 14 advanced dashboards for 6 teams (rail, flights, marketing, infra, accounting, support).
Analytics Engineering · Data Quality · Anomaly Detection · Streaming Analytics · Dashboarding · BI Migration · Metabase · Data Marts · ETL · DWH Design · Exasol · ClickHouse · MySQL · ElasticSearch · Prometheus

Product Analyst

May 2015 — Oct 2018
LinguaLeo · EdTech
  • Developed QlikView dashboards and reports for marketing, monetization, and product analytics. Refactored in-house analytics infrastructure from data collection and storage (Vertica, ClickHouse, Exasol) to reporting with Python and QlikView.
  • Integrated third-party systems: AppMetrica, Flurry, iTunes Connect, GA, Yandex.Metrica, App Annie, Amplitude. Generated hypotheses on user behavior and key metrics (LTV, ARPU, retention, ROMI), designed A/B test plans, and analyzed results using Bayesian and bootstrap methods.
  • Monitored metrics with early alerts and built ETL processes using Python and Ruby, optimizing queries in Redshift and Vertica while designing data storage for varied sources. Piloted data warehousing with ClickHouse and Exasol.
Product Analytics · Marketing Analytics · LTV · ARPU · Retention · Churn · ROMI · A/B Testing (Bayesian, Bootstrap) · Metric Monitoring & Alerting · Amplitude · AppMetrica · Google Analytics · QlikView · Python · ETL · Vertica · Redshift · ClickHouse

QlikView Developer

Jul 2013 — May 2015
BI Consult · Consulting
  • Delivered 4 QlikView / Qlik Sense projects for large enterprise clients in distribution, logistics, and FMCG.
  • Created an internal library of templates and design patterns, raising the design and structure of projects to a new level.
BI Development · Dashboard Design · Data Modeling · QlikView · Qlik Sense · Retail / Distribution / FMCG Analytics
PORTFOLIO

Less talk, more demos.

The open tiles are live: the first screen states the idea, the sibling screens let you drive it yourself. The locked ones are built, but not public.

Semantic layer client
OLAP-cube client

Illustrating a semantic layer client for tracking metrics. Slice and dice any metric in any split instantly, without any lag.

Stochastic model
What-if analysis

Example of a stochastic model of unit economics. Serves as a basis for scoring hypotheses.

Growth hypotheses
Hypothesis scoring

Growth-hypothesis scoring for real companies: expected value and value of information per hypothesis, priced through a causal model of the business.

Not public
JDBC · ClickHouse
ClickHouse semantic proxy

A JDBC gateway that translates semantic queries into native SQL — connecting any visualisation system to a custom semantic layer, and powering any BI with data from the OLAP cube.

Not public
Market research
Market intelligence

A body of research into different business niches. Competitive analysis.

Not public
SEMANTIC LAYER

The core
of a modern data platform.

The logic layer that canonically defines how metrics and dimensions are computed is the core of any mature analytics system. Too often this layer is smeared across visualisation tools — which makes it barely governable and lets it accumulate contradictions.

A semantic layer lifted out of the BI system lets you govern metric logic uniformly, use any visualisation tool, and make self-service and AI agents an effective way to access data.

Unlike a Context Layer, which contains only descriptions, a semantic layer answers semantic queries deterministically. This radically improves reliability and verifiability — which in text-to-SQL approaches is fundamentally limited.

DECISION INTELLIGENCE

BI was always meant to decide.

In modern usage, Business Intelligence often means nothing more than a visualisation system — which radically diverges from the original meaning and definition of BI as a decision-support system. That is why the industry needs to develop the notion of Decision Intelligence — a term to gather the methodologies of decision-making under one roof.

In b2c SaaS SMB, RICE is the usual way to score hypotheses and surface growth levers. Its grounding in data is very loose: half of RICE's inputs are subjective estimates. RICE is good where you need to align stakeholders and roughly size a backlog — but honest what-if analysis needs a structural causal model of the business (SCM). The news is that a technology previously affordable only to large corporations with dedicated data-science teams is, thanks to AI, becoming accessible and justified for SMBs and small product teams.