Saleema B.

Key Skills

Software

Label Studio

Top Subject Matter

Pharmaceutical Research & Drug Discovery

Chemistry

Pharmaceutical Patent Intelligence

Top Data Types

Text

Computer Code Programming

Document

Top Task Types

Entity (NER) Classification

Relationship

Classification

Question Answering

Text Summarization

Evaluation/Rating

Freelancer Overview

I have over 16 years of experience in pharmaceutical research, patent intelligence, scientific data curation, and technical documentation, with dedicated upskilling in cheminformatics, machine learning, and AI applications for chemistry. My core work has involved extracting, validating, and structuring complex chemical information from global patent literature and research publications including reaction pathways, multi-step synthesis routes, biological activity data (IC50, EC50, Ki), SAR information, Markush structures, and experimental procedures with >99% QC accuracy across 10,000+ records. This demands the same precision, consistency, and domain judgment that AI training and data labeling projects require. I have developed hands-on expertise in AI-assisted chemistry workflows, molecular data analysis, cheminformatics pipeline development (RDKit, PyTorch, AiZynthFinder), and scientific knowledge extraction. I am experienced in reviewing AI-generated chemistry content, verifying chemical information against authoritative sources such as Reaxys and CAS SciFinder, identifying factual and mechanistic inconsistencies, and providing domain-expert feedback to improve data quality and model performance. My combination of deep pharmaceutical chemistry knowledge, large-scale data curation experience, and applied ML skills makes me well-equipped to contribute to chemistry-focused AI training, annotation, and model evaluation projects.

Labeling Experience

Pharmaceutical Patent Data Curation & Scientific Annotation

TextEntity (NER) Classification

Extracted, validated, and structured chemical and biological data from 10,000+ pharmaceutical patents and scientific literature, including IC50, EC50, Ki bioactivity values, SAR information, Markush structures, and multi-step synthesis routes with >99% QC accuracy. Work involved domain-expert judgment to identify factual inconsistencies and ensure data quality — directly transferable to AI training and annotation tasks.

2021 - 2026

Markush Structure Analysis & Chemical Entity Annotation

ImageClassification

Analyzed and annotated complex Markush structures from pharmaceutical patents, decomposing generic chemical representations into specific compound classes, scaffold cores, and substituent variations. Identified and classified core scaffolds, R-group variations, and functional group patterns across large patent families. Applied deep domain knowledge to interpret chemical scope, flag ambiguous structures, and ensure consistent classification. This is directly analogous to chemical entity labeling and molecular annotation tasks required in AI chemistry training projects. Tools used: Reaxys, CAS SciFinder, STN, MDL ISIS/Draw, ChemDraw. Reviewed 10,000+ patent records with >99% QC accuracy across multi-jurisdiction patent families (USPTO, EPO, WO, JP).

2010 - 2021

Education

Work History