SB.
Let's talk
02Capabilities

A connected technical practice

From first schema
to final signal.

My work moves between the product surface and the infrastructure beneath it. Each layer is designed to be clear, observable and ready to evolve.

Interface
API + AI layer
Signal store
SignalDetails
01

Applied AI

RAG · Agentic AI · FAISS · Azure OpenAI · LLM deployment · YOLO · PyTorch · TensorFlow · model fine-tuning

02

Backend systems

Python · JavaScript · Django · Flask · NestJS · scalable REST APIs · RBAC · relational schemas

03

Data & automation

PostgreSQL · MySQL · MongoDB · Supabase · Python pipelines · UiPath · n8n · Power BI · Excel

04

Infrastructure

Docker · Podman · Linux · Git · CI/CD · RHCSA · practical service observability

Diagram / 02 · interactive

Technology decision map

The stack is selected around the problem, with clear boundaries between each responsibility. Select a node to inspect its role.

architecture mapstack topologystatus: nominal
System
Active layer / 01Product surface

Web and mobile interfaces create an intentional entry point.

connectedflow / stacknodes / 4click a layer to inspect
PipelineShipping rail
CI/CD / delivery rail

From commit to healthy in one track.

Every push flows through a repeatable, observable pipeline. Quality gates and automatic rollback keep the ship moving without breaking production.

pipeline topologyenvironment: production6 stages · on rails
pipeline / 01 · CommitSource controlpush · PR · review

The pipeline starts with a push. Branches are protected, PRs are reviewed and every merge produces a single source of truth for the build ahead.

  • GitHub
  • Git
  • branch protection
event
push → main
state
passed
artifact
Git
pipeline statusgreen · main branchtotal runtime25.4sdeploys this week14click a stage to inspect
DataData practice
Data science / practice rail

From raw noise to useful signal.

An applied data workflow: collect, clean, store, explore, model and serve — with quality gates at every step so the signal stays trustworthy.

pipeline topologyenvironment: production6 stages · on rails
pipeline / 01 · IngestCollectstream · batch · sources

Raw material arrives from APIs, files, databases and automation runs. Python pipelines and UiPath pull whatever the system needs, on a schedule or on demand.

  • Python pipelines
  • UiPath
  • n8n
  • webhooks
sources
stream + batch
state
synced
artifact
UiPath
pipeline statusgreen · main branchtotal runtime25.4sdeploys this week14click a stage to inspect
InfraCompute
compute optionscloud ⇄ privateright tool · right place
Public cloudelastic · hands-off
Computeserverless + autoscale
AI / LLMhosted models · managed RAG
Datamanaged Postgres + vector
Networkglobal CDN · edge
Opspay-as-you-go · zero infra
Private computeowned · local · private
Computebare metal + containers
AI / LLMlocal / edge models · data stays home
Dataself-hosted Postgres + files
NetworkLAN · air-gapped option
Opsfull control · fixed cost
cost
pay-as-you-gocapex · predictable
latency
network hopnear-native
control
managedfull root
scale
elasticbounded

Most systems run a hybrid: heavy lifting in the cloud, sensitive data and low-latency work on private compute. Right tool, right place.

Continue exploring

Next signal