Hamza Mahmood · imagine-talent

Email

hamzamahmood165@gmail.com

Phone

923190476259

https://www.linkedin.com/in/hamza-mahmood-57447024a/

GitHub

—

Academic

Program

BSCS

CGPA

—

Year

2022

Education

SEECS

Address

252-J TECH TOWN ,SATYANA ROAD, FAISALABAD. , Faisalabad , Pakistan

DOB

—

Career

Current role

—

Target role

—

Skills

Deep Learning, Computer Vision, LLMs, RAG, AI Agents, Chroma, Vector Databases, Knowledge Distillation, Pruning, MOE, CKA Similarity, DocVQA, FUNSD, ANLS, Python, Automation, RSS, YouTube API

Verbatim text

The exact text the LLM saw on the page (or the booklet text from the old import). This is what powers semantic search.

Hamza Mahmood
Cell: 923190476259 | Email: hamzamahmood165@gmail.com
LinkedIn: https://www.linkedin.com/in/hamza-mahmood-57447024a/
Address: 252-J TECH TOWN ,SATYANA ROAD, FAISALABAD. , Faisalabad , Pakistan
PROFESSIONAL PROFILE
I am a final year student at NUST pursuing my bachelors in Computer Science.
I have expertise in deep learning, computer vision , LLMs, RAG and AI agents,
EDUCATION
BSCS
SEECS , Islamabd , 2.97 (2022)
INTERNSHIP EXPERIENCE
PIpeline Scientists
26-Jun-2025 - 26-Sep-2025
• Developed a context aware ”Reddit Brand Ambassador” for ColdSend.pro that helps people and pitches the product in relevant
threads. • Developed a Reddit blog posting automation connected with the RSS feed of product’s website. • Developed Youtube
comment posting automations to increase the reach of the product.
FINAL YEAR PROJECT
LiteDoc: Distilling large document models into task specific encoders.
* Working on the distillation of Deepseek VL-2 as part of my FYP. ∗ Pruning Experts present in the MOE layer based on the CKA
similarity of the experts then using knowledge distillation to increase the pruned model’s performance. ∗ “Using DocVQA and
FUNSD to test the model’s performance and ANLS score as an evaluation metrics.
TECHNICAL EXPERTISE
LLMs
∗ Developed a Local RAG-Based Financial Education Chatbot: Implemented a Retrieval-Augmented Generation (RAG) system
utilizing vector databases (Chroma) and Large Language Models (LLMs) to provide instant, contextually relevant answers to ﬁnancial
queries based on curated educational content. ∗ Optimized ...

AI enrichment

Hamza Mahmood is a final-year BSCS student at NUST with expertise in deep learning, computer vision, and LLMs. He has internship experience in AI agent development and is currently working on a final year project involving model distillation and pruning.

Skills (AI)

["Deep Learning", "Computer Vision", "LLMs", "RAG", "AI Agents", "Knowledge Distillation", "Model Pruning", "Vector Databases", "Python"]

Status: ai_done

Provenance

Source file: SEECS - Computer Science-2026.pdf
From job #258 page 103
Created: 1778167261