Anton O.

Project overview:

A telecom company had large volumes of messaging and campaign data but no practical way to use it for targeting or personalization. The goal was to build production ML systems that could directly improve campaign performance.

Responsibilities:

• Collected requirements directly from marketing and operations teams
• Built NLP pipelines for message classification and user profiling.
• Developed a recommender system to match campaigns to likely recipients.
• Migrated data processing and ML workloads to GCP.
• Built simple internal tools so non-technical users could manage campaigns.

Achievements:

Several ML-powered features were deployed to production and used by marketing teams. Campaign revenue increased substantially without increasing traffic costs.

Project overview:

A consumer-facing live streaming platform needed to improve real-time content recommendations under strict latency constraints. Traditional offline-trained models were slow to adapt to user behavior and changing content dynamics. The goal was to design a system that could learn continuously from live user interactions.

Responsibilities:

• Designed the end-to-end recommendation architecture combining offline training and online learning.
• Defined reward signals based on user interactions (watch time, skips, engagement events).
• Built a real-time inference service with tight latency budgets.
• Implemented safeguards to prevent feedback loops and degraded user experience during exploration.
• Worked closely with product and backend teams to integrate the model into the live serving stack.

Achievements:

A reinforcement learning–based recommendation system was deployed to production, adapting recommendations in near real time based on user feedback signals. The system improved engagement metrics while remaining stable under high request rates.

Project overview:

A crypto investment fund needed to run large-scale backtests over long historical time ranges. An initial cloud-first design using managed services proved too expensive once realistic workloads were tested.

Responsibilities:

• Designed the initial research pipeline for time-series analysis.
• Identified cost blow-ups caused by excessive orchestration and event fan-out.
• Stopped the rollout early, explained the cost drivers to stakeholders, and proposed alternatives.
• Rebuilt the execution model using simpler scheduling and dedicated compute.
• Rewrote infrastructure-as-code and data flow to remove unnecessary stages.

Achievements:

The architecture was simplified and partially moved off managed cloud services. The final solution delivered the required compute throughput at a small fraction of the original projected cost, making the research financially viable.