About

I’m Cong Peng, an Engineer / Architect / Builder based in Stockholm. I build agentic AI applications and the cloud data & AI platforms behind them. I lead architecture and hands-on implementation from MVP to production.

Currently building Actorise.

Some consulting projects

Agent AI apps

Built multi-agent data analytics solution where a supervisor plans tasks and sub-agents iteratively generate/execute SQL and synthesize reports
Shipped production RAG as a reusable reference implementation, combining context engineering, retrieval design, evaluation/guardrails, and observability
Delivered LLM-powered workflow automation integrated via async APIs and secure enterprise integration patterns

AgentOps / MLOps / DataOps in the cloud

Designed platform patterns for fast provisioning, CI/CD, governance, and operational guardrails for AI apps across teams
Established end-to-end lifecycle foundations to operationalize data, models and agents with reusable pipelines and quality monitoring
Extended regulated enterprise platforms to adopt GenAI safely, focusing on architecture, security, and delivery standards

Machine Learning

Developed classic ML models end-to-end with feature engineering, scalable training, hyperparameter search, reproducible experiment tracking, production deployment and monitoring

Data engineering

Architected and implemented data pipelines with medallion architecture to serve analytics and ML

Some side projects

LinkedInfo.co

The Web should be an open web. All the informations published on the Web are meant to be shared, share through links by search engines, rss, social networks, etc. This site is yet another method that tries to link all the informations (but starts with only technical articles on LinkedInfo) and share them.

The original idea of this side project is to utilize Semantic Web technologies and Machine learning to link the informations. Noble ambition shall start from basic, it needs to be improved little by little.

Links: linkedinfo.co (inactive) · legacy page

Text Analysis Service

A text analysis service for technical articles includes topic identification and keywords extraction. The model of topic identification uses a pre-trained BERT and fine-tuned on the dataset of LinkedInfo.co and questions on Stack Overflow.

Links: demo · backup demo · legacy page

XGBoost.swift

The first Swift interface for XGBoost, which is an optimized distributed gradient boosting library implements Machine Learning algorithms under Gradient Boosting framework.

Links: GitHub · legacy page

Skin Lesion Classifier

A skin lesion classifier that uses a deep neural network trained on the HAM10000 dataset. An implementation of the ISIC challenge 2018 task 3.