MOD / CASES

Selected products

Complex products designed and delivered end-to-end: AI SaaS, HRTech, data platforms, e-commerce, Web3 and applied machine learning.

05 CASE FILES

The UM ecosystem

Products with one engineering culture

Case studies

Shown: 5 / 5
Filter:

AvtoUM

Year2026
avtoum.ru · AI salesperson · SaaS

AI salesperson: consults the client and closes to a lead — 24/7 on site, VK and Avito

One core connects three channels, AI conversations, a lightweight CRM and live Pulse telemetry. The owner sees incoming requests in real time and can step in when a human is needed.

LLMAI-агентCRMRealtime
Key figures
34 687
dialogs/mo
21 67062%
leads
Dialogs → leads
Dialogs / mo34 687
Leads21 670
27
Business niches
63
Companies
3
Channels
1.6M
API tokens

SvoyUM

Year2026
svoyum.ru · Android · product architecture · privacy

A local-first personal progress system: one event model for tasks, projects, learning, movement, books, habits and finance

Local events form the product core: an action becomes a fact, then an explainable insight. The database stays on-device; migrations and backup preserve data as the product evolves.

AndroidLocal-firstProduct architecturePrivacyMigrations
Key figures
ANDROID
released product
LOCAL
personal data on-device
From action to insight
  1. 01
    Action
  2. 02
    Core
  3. 03
    Insight
  4. 04
    Privacy
APK
Status
LOCAL-FIRST
Core
MIGRATIONS
Data
INTERNAL BETA
Cloud

4iTX

Year2026
4itx.ru · HRTech platform · Data · AI

HRTech platform for Russia’s deep-tech talent market: AI-classified vacancies and market data

Jobs from multiple sources are cleaned and normalized, then AI-classified by direction and grade. Users get filters for skills, work format and salary instead of scattered listings.

HRTechDataLLMWeb3Cloud
Key figures
12 000+
vacancies
5
directions
Vacancy pipeline
  1. 01
    Sources
  2. 02
    Parse / clean
  3. 03
    AI classification
  4. 04
    Storefront
3
Sources
~12 000
Indexed
~7 400
With salary

UMZARUMZAR

Year2025
ФИПС · гос. регистрация · Data Science · taxonomy · verification

Scientific taxonomy of world commercial timber (WCT)

One trade name may refer to different species. UMZAR brings names and properties into one model: taxonomy, unique codes and a 25-cell matrix make records searchable, comparable and verifiable. The model supports trade, wood processing, customs and research.

Data ScienceТаксономияМатрицаVerification
Key figures
186
verified species
459
timber subspecies
Classifier matrix · 5 × 5
Density, kg/m³ →
Durability →
P1
P2
P3
P4
P5
D1–P1
D1–P2
D1–P3
D1–P4
D1–P5
D2–P1
D2–P2
D2–P3
D2–P4
D2–P5
D3–P1
D3–P2
D3–P3
D3–P4
D3–P5
D4–P1
D4–P2
D4–P3
D4–P4
D4–P5
D5–P1
D5–P2
D5–P3
D5–P4
D5–P5
P1 <500P2 500–650P3 650–800P4 800–950P5 >950kg/m³
D1 Non-durable · <5 yearsD2 Slightly durable · 5–10 yearsD3 Moderately durable · 10–15 yearsD4 Durable · 15–25 yearsD5 Highly durable · >25 years
Матрица
Method
Data Science
Approach

MEDCYB

Year2024
СПбГПМУ · Machine Learning · medical · research

ML recognition of gestational diabetes predictors in pregnant women

The research workflow covers data-quality checks, EDA, feature preparation and validation of classification approaches. Findings are reviewed with clinicians; the model does not replace clinical judgement.

Machine LearningData ScienceHealthcareResearch
Key figures
15 000+
clinical observations
~85
source parameters
Research pipeline
  1. 01
    Dataset
  2. 02
    Preparation
  3. 03
    Feature selection
  4. 04
    Model
ML
Method
ГСД
Target