Looks like you have JavaScript disabled. For the full Twine experience, you will need to re-enable it.

Hi, I'm Milos Nikolic, a senior AI/ML engineer focused on agentic AI, RAG, and scalable ML platforms. I design and ship production-grade solutions that boost task success, reduce hallucinations, and scale responsibly across web-scale and enterprise contexts. I enjoy turning ambitious ML ideas into measurable business outcomes, building robust data and model pipelines, and mentoring teams to operate reliable ML systems. I thrive on cross-functional collaboration with product and engineering teams and value responsible AI practices.…Hi, I'm Milos Nikolic, a senior AI/ML engineer focused on agentic AI, RAG, and scalable ML platforms. I design and ship production-grade solutions that boost task success, reduce hallucinations, and scale responsibly across web-scale and enterprise contexts. I enjoy turning ambitious ML ideas into measurable business outcomes, building robust data and model pipelines, and mentoring teams to operate reliable ML systems. I thrive on cross-functional collaboration with product and engineering teams and value responsible AI practices.

Milos Nikolic

AI Engineer, Data Scientist, AI Developer, +2





Hi, I'm Milos Nikolic, a senior AI/ML engineer focused on agentic AI, RAG, and scalable ML platforms. I design and ship production-grade solutions that boost task success, reduce hallucinations, and scale responsibly across web-scale and enterprise contexts. I enjoy turning ambitious ML ideas into measurable business outcomes, building robust data and model pipelines, and mentoring teams to operate reliable ML systems. I thrive on cross-functional collaboration with product and engineering teams and value responsible AI practices.…Hi, I'm Milos Nikolic, a senior AI/ML engineer focused on agentic AI, RAG, and scalable ML platforms. I design and ship production-grade solutions that boost task success, reduce hallucinations, and scale responsibly across web-scale and enterprise contexts. I enjoy turning ambitious ML ideas into measurable business outcomes, building robust data and model pipelines, and mentoring teams to operate reliable ML systems. I thrive on cross-functional collaboration with product and engineering teams and value responsible AI practices.

Available to hire

Hi, I’m Milos Nikolic, a senior AI/ML engineer focused on agentic AI, RAG, and scalable ML platforms. I design and ship production-grade solutions that boost task success, reduce hallucinations, and scale responsibly across web-scale and enterprise contexts.

I enjoy turning ambitious ML ideas into measurable business outcomes, building robust data and model pipelines, and mentoring teams to operate reliable ML systems. I thrive on cross-functional collaboration with product and engineering teams and value responsible AI practices.

Skills

Experience Level

Expert

Expert

Expert

Expert

Expert

Expert

Expert

Intermediate

Work Experience

Senior AI/ML Engineer at META AI

July 1, 2025 - October 15, 2025

Architected a production multi-agent system (Planner -> Actioner -> Executor -> Feedback -> Evaluator) with self-correcting loops; built RAG pipeline with hybrid search and LLM summarization; improved task success rate to 85%, reduced hallucinations by 35%, and increased retrieval precision by 85%. Delivered a web-scale ranking platform with TensorFlow/TFX on Vertex AI, Dataflow, and BigQuery; implemented privacy-preserving personalization (Federated Learning + DP) and established standard experimentation suites. Led reliability program with 99.9% inference SLO and mentored 6 DS/ML Engs, partnering with 4 product teams.

Data Scientist at Databricks

December 1, 2022 - October 15, 2025

Productionized churn & propensity models on SageMaker with model registry, blue-green deployment, and drift monitoring; reduced churn by 22% across three pilot cohorts (~45k users). Built a real-time lakehouse data plane ingesting 10+ TB/day with 1.8k QPS inference using Spark/Kafka; delivered LATAM-regulated templates and improved observability to reduce time-to-prod by 6 hours (from ~3 weeks) and cut infra costs by 18%. Implemented feature lineage and calibration for model governance.

Principal ML Consultant at Capgemini Invent

November 1, 2019 - October 15, 2025

Retail forecasting with hybrid classical + deep learning models and hierarchical reconciliation, reducing stockouts by 17% and holding costs by 11% across 220 stores. Last-mile optimization with learning-to-rank and heuristic search reduced last-mile costs by 9% and improved on-time delivery by 6% for 50k+ deliveries/week. Developed responsible-AI guardrails, bias scanning, and model cards to streamline client reviews and post-deployment governance.

Machine Learning Engineer at Holycode

July 1, 2017 - October 15, 2025

Built an enterprise labeling platform (CV/NLP) with Flask/React to train/retrain models; reduced dataset turnaround by 50% (4 weeks to 2 weeks) for a tier-1 bank & public-sector client. Deployed anti-spoofing and error-monitoring models, achieving ~23% QoQ improvements in incident handling, and implemented serverless identity management to shorten access ticket resolution times by ~30%.