Recursive Data

Senior Data Scientist

Turing • Full-time

Apr 2024 - Jan 2026 • 1 yr 9 mos

Spearheaded the development and optimization of a cutting-edge LLM-based system designed to train data analysts by integrating intelligent query understanding and adaptive learning pathways. Engineered robust natural language processing (NLP) pipelines to transform unstructured business queries into actionable data insights using state-of-the-art language models. Collaborated with cross-functional teams to fine-tune LLM models on domain-specific datasets, significantly improving response accuracy and contextual relevance for real-world data analytics tasks. Designed and executed prompt engineering strategies to enhance model performance, tailored to specific analytical intents and workflows in a multimodal AI pipeline. Led evaluation of generated outputs across different models using custom scoring metrics and human-in-the-loop feedback systems to determine the most contextually accurate responses.

Senior Full stack web and AI developer

Zaytrics (Pvt.) Limited • Full-time

Dec 2020 - Apr 2024 • 3 yrs 4 mos

Developed intelligent PDF processing tools in Python to extract highlighted text and implemented header/footer detection and removal algorithms for cleaner document parsing. Built and trained machine learning models for text classification using techniques such as TF-IDF, Word2Vec, SVM, and Random Forest, tailored for domain-specific content. Implemented extractive summarization using Latent Semantic Analysis (LSA) and abstractive summarization with Seq2Seq (Encoder-Decoder) models trained on the CNN/DailyMail dataset. Designed and deployed RESTful APIs using Python (Flask) and Node.js (Express), integrating with React.js frontends and custom stored procedures for generating dynamic CSV reports. Created secure and scalable cloud deployments using AWS EC2 and FastAPI (Ray Cluster) for high-performance AI inference workloads.