Hi, I’m Tomer.

Data science & AI leader. Most recently led a group of 15 at Salesforce shipping production generative AI over large-scale enterprise data. Open to remote roles, full or part time.

Years in Data & AI12+
Years leading teams6+
Group led15 people
US patents5
Tomer Parash Mapa
Óbidos, Portugal

A bit about me.

Background

I'm a seasoned algorithm and data science leader. I pair deep technical expertise with hands-on people leadership: building and running teams, hiring and mentoring, and setting long-term R&D strategy. I've done this across large enterprises, mid-size companies and high-growth startups.

For the last several years I was at Salesforce, leading a group of 15 talented data scientists and developers at the forefront of how organizations use GenAI to interact with large-scale structured and semi-structured data. We specialized in the non-trivial problems, where meticulous data science and a high accuracy bar make the difference between a demo and a product that delivers real value. I led solutions like Einstein Segment Creation and the grounding layer behind Einstein Copilot from inception all the way to production.

Away from the keyboard I'm married to Dorin and the father of Alma and Omri. We spent the past year on a planned family sabbatical, traveling the world together. Now I'm back, recharged, and looking for my next challenge: a remote role, full or part time.

Where I’ve done the work.

salesforce
Tel Aviv, Israel · 2020–2026

Senior Manager, Data Science, Generative AI

2023–2026
  • Led 15 data scientists and developers building GenAI for the way enterprises interact with large-scale structured and semi-structured data, taking on non-trivial problems with a high accuracy bar.
  • Shipped Einstein Segment Creation: marketers define audience segments in natural language over any data lake, with no prior training on the organization’s data or schema.
  • Delivered natural-language SQL Q&A over vast, un-organized enterprise tables, combining hybrid semantic and lexical search with LLMs.
  • Built a RAG-over-metadata service returning the most relevant tables, columns and values for any utterance. This service is the grounding layer behind Einstein Copilot and the two solutions above.
  • Engineered on direct provider SDKs and an in-house orchestration layer, with an eval harness and golden datasets gating every release.
  • Owned hiring, mentoring and long-term R&D strategy for the group.

Data Science Manager, Service Cloud AI

2020–2023
  • Led five data scientists and ML engineers building AI for service and contact centers: workforce management, capacity planning and time-series forecasting of demand.
  • Owned delivery end to end, from business framing to production.
ClickSoftware
Acquired by Salesforce · 2017–2019

Data Science Manager

2019
  • Led three data scientists on ML over tabular data for field-service management, owning technical direction, mentoring and the product roadmap.

Data Scientist

2017–2019
  • Built ML products end to end: classical and NLP feature engineering from structured and unstructured sources, model tuning, and real-time on-demand prediction in production.
Earlier
2012–2017

Data Scientist · TGS

2016–2017
  • Recommenders, anomaly detection, time-series and classification.

BI Developer · VolcanoData (Teddy Sagi Group)

2015–2016
  • E-commerce BI on Hadoop; Tableau, Pentaho, MySQL, R.

Lecturer & Teaching Assistant · Ruppin & BGU

2012–2015
  • OOP, Quality Control and Simulation.

Patents & published research.

US Patent · Granted · US-11823104-B2

Systems and Methods for Scheduling Connected Device

Granted by the USPTO, 2023.

US Patent · US-20250147987-A1

Database Record Creation from Text-Based Interactions

Published by the USPTO, 2025.

US Patent · US-20200210939-A1

Systems and Methods for Assigning Tasks Based on Real-Time Conditions

Published by the USPTO, 2020.

US Patent · US-20200210963-A1

Methods and Systems for Appointment Booking

Published by the USPTO, 2020.

US Patent · US-20200210933-A1

Systems and Methods for Scheduling Connected Device

Published by the USPTO, 2020.

Paper · IJCAI 2015

Applying Max-Sum to Asymmetric DCOP

International Joint Conference on Artificial Intelligence.

Paper · AAMAS 2017

Balancing Exploration and Exploitation in Incomplete Min/Max-Sum Inference for DCOP

International Conference on Autonomous Agents and Multi-Agent Systems.

Five published US patents · Two peer-reviewed papers

Education.

2012–2014

M.Sc. Industrial Engineering & Management, major in Artificial Intelligence

Ben-Gurion University · GPA 87.6 · Excellent-students program

Thesis: heuristics for distributed incomplete algorithms in multi-agent systems.

2009–2013

B.Sc. Industrial Engineering & Management

Ben-Gurion University · Graduated with distinction · GPA 86.2

What I work with.

GenAI

LLM applications, RAG, hybrid semantic & lexical retrieval, reranking, NL-to-SQL, context engineering.

LLM Evaluation

Eval harnesses, golden datasets, accuracy benchmarking, regression gates in CI.

Data Science

Time-series forecasting, tabular classification & regression, NLP, anomaly detection, recommenders.

Tech

Python, SQL, vector DBs, AWS cloud architecture, production ML pipelines.

Leadership

Team building, hiring, mentoring, R&D strategy, cross-functional delivery.

Languages

Native in Hebrew, fluent in English.

Ready to build together?

Always happy to talk. Don't be a stranger!

Or drop me a line right here