Abheet Aggarwal

Product Engineer

Bengaluru, Karnataka, India8 yrs 6 mos experience

Most Likely To SwitchHighly Stable

Key Highlights

7+ years of experience in data science.
Expert in machine learning and deep learning algorithms.
Ranked 30th in global apparel attributes extraction competition.

Stackforce AI infers this person is a Data Science expert with a focus on NLP and Computer Vision.

Contact

Skills

Core Skills

Data ScienceLarge Language Models (llm)Creative Problem SolvingDeep Learning

Other Skills

Responsible AIConsumer BehaviorLSTMSequence to sequence ModelsAttention MechanismRogue ScoreTopic ModellingCNNTransfer LearningHadoopApplied MathematicsMachine LearningPythonArtificial Neural NetworksTime Series Analysis

About

IITK Graduate-Lead Data Scientist with 7+ years of experience in building statistical models using huge datasets in Media, CPG & academic domains. Proficient in machine learning & deep learning algorithms, python programming & exploratory data analysis. NLP product expert, enthusiast for solving difficult problems

Experience

Mastercard

Lead Data Scientist

Oct 2025 – Present · 5 mos · Pune District, Maharashtra, India · On-site

Turbohire

3 roles

Chief Data Scientist

Promoted

May 2022 – Oct 2025 · 3 yrs 5 mos

Responsible AICreative Problem SolvingData ScienceLarge Language Models (LLM)

Lead Data Scientist

Jan 2022 – Jun 2022 · 5 mos

Senior Data Scientist

Mar 2021 – Jan 2022 · 10 mos

Fractal

2 roles

Senior Data Scientist

Feb 2019 – Mar 2021 · 2 yrs 1 mo

Consumer BehaviorData Science

Senior Data Scientist

Sep 2017 – Mar 2021 · 3 yrs 6 mos

TEXT SUMMARIZATION :
Effectively summarizing the long documents at scale can speed up consumption of information
effectively.
Evaluated Abstractive (LSTM – Sequence to sequence Models with attention) & Extractive algorithms for the task.
Different metrics were researched to find a metric which can capture the meaning of text more
accurately, like rogue score and topic modelling.
Attributes Extraction of Apparel:
Identifying the attributes of the apparel from an image . Data provided has 8 categories, each
representing a clothing type. Each category is further broken down into labels defining it in terms of design or length
Used Deep learning - CNN to solve the problem. Many different models were experimented for
this.
Transfer learning was used for the final solution. This helps in creating the initial base features and
avoids training the model from scratch when you have limited data and computational resources.
The competition was hosted at the Alibaba cloud competitions site: Tianchi. Ranked 30th out of
2995 participants across the world