Looking for a job

Kaveh Momeni

@kavehmomeni

Message

AI Tech Lead at Chaharsotoon Engineers who cut project-risk detection lag from over 10 days to under 24 hours.

Germany

Message

What I'm looking for

I'm seeking an AI Tech Lead position in Berlin where I can apply my expertise in AI architecture design and cloud services to enhance AI capabilities and deliver measurable business value.

At Chaharsotoon Engineers, I developed a production project-risk classifier that reduced detection lag from over 10 days to under 24 hours and cut manual review by about 80%. I also introduced NLP-based project-scope and change-order features that improved pricing-prediction error by about 18%.

With DISPOTIF, I took an AI film pre-production planning product from concept to a live bilingual English/Persian app in about three months. I also built an evaluation-gated ML runtime for the Numerai Tournament, with human approvals, audit trails, monitoring, and rollback controls.

Experience

Work history, roles, and key accomplishments

Data Scientist / AI Project Lead

Chaharsotoon Engineers

Nov 2022 - Dec 2025 (3 years 1 month)

Developed a production project-risk classifier using transformer-based NLP features and LightGBM/XGBoost stacking, reducing detection lag from over 10 days to under 24 hours and cutting manual review by ~80%. Improved pricing-prediction error by ~18% with NLP-based features and deployed multiple AI workflows with IAM-aware access controls.

Python NLP LightGBM XGBoost Optuna SHAP Airflow PostgreSQL Power BI MLOps Jira Docker Kubernetes fastAPI CI CD Metabase

Data Scientist

Snappstore

Nov 2023 - Sep 2024 (10 months)

Developed a weekly demand forecaster using a hybrid ARIMA + LSTM neural network, achieving 10-15% MAPE and reducing stock-outs by ~12%. Deployed the model as a FastAPI microservice on Docker and Kubernetes with Metabase monitoring dashboards.

Python ARIMA LSTM fastAPI Docker Kubernetes MLOps Metabase CRM OMS POS

R&D / Data Scientist

Kongereh Platform

Feb 2020 - Oct 2022 (2 years 8 months)

Planned hybrid recommendation models for a 10,000+ SKU catalogue and built TF-IDF + SVM intent classification for procurement requests. Reduced manual reporting from ~10 to ~2 hours/week and decreased report turnaround from ~2 days to ~3 hours.

Python SQL TF IDF SVM PostgreSQL MLOps ETL Dashboarding