← Back to cohort

Harras Mansoor

FAST · 2025
Email
harrasmansoor3@gmail.com
Phone
+923235055205
LinkedIn
https://www.linkedin.com/in/harras-mansoor/
GitHub

Academic

Program
CGPA
Year
2025
Education
Address
DOB

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.
Harras Mansoor
+923235055205, harrasmansoor3@gmail.com
Ghaziabad, Rawalpindi.
Linkedln: https://www.linkedin.com/in/harras-mansoor/
ineerin
inal Pro ·ect: Caselink All-in-One software solution for law firms
aselink is an Al-powered all-in-one legal tech platform that streamlines legal research, case management, document
utomation, and client communication for law firms.
emester Pro ·ects:
allucination free Speech-to-text [Python]:
educed hallucinations in real-time Speech-to-Text solutions
ulti-threaded Neural Network training [C++):
eveloped a dummy neural network tailored for a parallel processing setting within multicore systems
lockchain for Al model training [Go]:
ecentralized blockchain framework enabling collaborative Al model training, verification, and sharing with
ransparency and trust.
vent Management System [C#]:
eveloped an event management system in c#
imetable Scheduling Problem [Python]:
olved Timetable schedulin
roblem efficient! usin Genetic Al orithm
I INTERN, Artificial Intelligence and Machine Learning Lab (AIM LAB)
worked on the Research and development of a python library(Urduistics) for Natural
anguage Processing. I built various tools which can help developers and researchers to
ork on NLP problems in our local language "URDU" and also "Roman Urdu".
I INTERN, AIO
worked on Text and Audio based Al solutions. I Developed techniques for accurate
peech-to-Text and Worked on reducing hallucinations in Speech-to-Text.
kills & Tools
rofessional Skills
July 2022 - August 2022
June 2024 - August 2024
thon, Tensorflow, P torch, Cuda, C++, C, C#, TensorRT, Triton, Lan chain, Lan Gra h, crewAI
ublished Research Paper (Mitigating Hallucinations in Speech-to-text systems) in IEEE 4th ICMI
WS Scholar: Received 2x scholarship by AWS for two Nanodegrees
resident of Microsoft Learn Student Ambassadors FAST NUCES ISB Cha ter
vidia certification in Model Parallelism: Building and Deploying Large Neural Networks
vidia certification in Com uter Vision for industrial Ins ection, Dee Learn in S ecialization b Coursera
ctivities
osted a Hackathon in Nascon '23 and a Panel Discussion in Nascon '24
unner up, ldeathon. Second Runner up, DataQuest
la ing Cricket and table tennis
FAST NUCES ISLAMABAD CAMPUS

AI enrichment

Harras Mansoor is a student at FAST NUCES with internship experience in AI and NLP, focusing on speech-to-text optimization and Urdu language processing. He has published research on mitigating hallucinations in AI systems and holds certifications in deep learning and model parallelism.
Skills (AI)
["Python", "C++", "C#", "Go", "TensorFlow", "PyTorch", "NLP", "Speech-to-Text", "Genetic Algorithms", "Blockchain", "CUDA", "TensorRT", "LangChain", "crewAI"]
Status: ai_done
Provenance
Source file:
Created: 1777724106