Delivery Lead, AI & ML
Denitsa Katsarska
Owns the AI and machine-learning side of delivery: where a model earns its place in a process, and how its answers are checked.
- 6+ years in intelligent automation and AI/ML delivery
- Sofia, Bulgaria
- Bulgarian (native), English (fluent), Italian (conversational)
Expertise
- Applied AI & ML in process
- ML lifecycle & deployment
- Model evaluation & thresholds
- Document & communication understanding
- Low-code automation platforms
- Compliance-heavy automation
- Operating models & COE
- Change & adoption
Experience
Delivery Lead, AI & ML
Emerant, Now
Senior Consultant
Global consulting firm, 2024 - Now
Consultant
Global consulting firm, 2022 - 2023
Engineer, Customer Service
Enterprise software support, 2020 - 2022
Sales Executive
Cyber security reseller, 2018 - 2019
Legal Adviser
Private enforcement practice, 2014 - 2018, 2019 - 2020
Strategic priorities
Where a model belongs
And, more often, where it does not. Most of what gets called AI is a decision someone used to make quickly and is now asked to make at volume.
Measured before relied on
An evaluation set before a prompt, and a failure rate that can be shown rather than asserted.
Uncertainty routed, not guessed
Thresholds that send the unclear cases to a human instead of forcing an answer.
Vision
You will see the model's failure rate before you depend on it, and a human path for every case it should not decide.
Talk to Denitsa
Write directly, or book a first conversation with the team.