ST Engineering, GTO, AI.R STC, GenAI

What we're building

Explore our latest GenAI projects

The collection

Explore what's possible.

From first experiment to live pilot, each project is easy to explore, understand, and share.

3 projects

Updated as we build
Document AI

Tender Review System

AI-powered tender analysis platform that turns complex document sets into clear, actionable insight.

Powered by QwenTender intelligence, grounded in your documents.
Learn more
Voice AI

eServ AI Call Center

AI-powered call-center analytics platform that transforms call audio into actionable insight.

Powered by Gemma 4Call intelligence, on your infrastructure.
Learn more
OSINT

Geospatial Intelligence - The Nexus

Real-time situational-awareness platform that fuses open-source and operational data into one live, time-aware map.

Powered by Gemma 4 and xAIConnected intelligence for a live operating picture.
Learn more

How we build

Small experiments. Serious outcomes.

Every prototype starts with useful work, not technology for its own sake.

  1. 01

    Find the friction.

    Start with a real human problem worth solving.

  2. 02

    Make it tangible.

    Turn an idea into something people can try quickly.

  3. 03

    Learn in the open.

    Demo, measure, listen, and build the next iteration.

Our operating model

One team. Three connected strengths.

We bring together secure compute, efficient models, and rapid delivery as one practical GenAI capability.

Diagram of the GenAI Team operating model connecting secure edge and compute, efficient LLMs, and rapid delivery through an orchestration and automation pipeline.
View full-size diagram
Read the operating model

GenAI Team — Orchestrating Intelligence. Delivering Impact.

Edge & Compute

Secure. Sovereign. Scalable.

  • Air-Gapped / Sovereign AI
  • On-Prem Infrastructure
  • Edge Deployment
  • Real-Time Inference
  • Secure by Design
  • NVIDIA DGX · Apple Mac Studio

Deploy anywhere.

Efficient LLMs

Smaller. Faster. Smarter.

Model compression lifecycle: 70B → 7B → 3B → Edge.

  • Quantization
  • Pruning
  • Distillation
  • Tensor Compression
  • Efficient Inference
Deliver Rapidly
  1. Plan — AI Agent
  2. Build — AI Coding Copilot
  3. Review — AI Code Reviewer (Quality, Security, Best Practices)
  4. Test — AI Test Agent
  5. Deploy — Automated Delivery (Build, Test, Security Scan, Deploy)
  6. Observe — Feedback Loop

Orchestration & Automation Pipeline: Agents → Tools → APIs → Data → Models → Deployments.

GenAI team

We Are a Small And Agile Team

  • Michal Polanowski, PhD, MBA

    Portrait of Michal Polanowski.

    More than three years with ST Engineering. Focused on the efficient use of AI across software and hardware, including AI agents and AI personas. Also consults business units on GenAI adoption and builds proofs of concept for them.

  • Shee Jia Wei, MEng

    Portrait of Shee Jia Wei.

    More than one year with ST Engineering. Focused in hardware-efficient AI infrastructure, self-hosted LLM deployment, distributed computing, and value added AI solutions. Experienced in leveraging NVIDIA DGX Spark, Apple Silicon, and heterogeneous computing environments to deliver scalable AI capabilities without relying on conventional high-cost data-center GPU infrastructure. Strong background in production computer vision, MLOps, and AI system performance engineering.

  • Pranav

    Portrait of Pranav.

    AI engineer for the past 8 months at ST Engineering, having been a Data Scientist previously. I love creating GenAI harnesses with the aim of simplifying workflows

  • Yang Ying, MTech

    Portrait of Yang Ying.

    Senior AI Engineer with 8+ years of experience in research and industry, specializing in computer vision, multimodal learning, and generative AI, with a focus on developing innovative and practical AI solutions.

  • Shubhanker

    Portrait of Shubhanker.

    An overqualified AI intern with strong hands-on technical skills and a builder’s mindset—turning complex ideas into practical GenAI solutions. Passionate about solving real-world problems, simplifying workflows, and making AI useful in everyday work.