Harshit Agrawal

DS2 @ Amazon | LLMs | Deep Learning | IIT Roorkee

Bengaluru, Karnataka, India

About

• Specialization: Gen AI (Large Language Models), Deep Learning & Deep Reinforcement Learning. • Proven track record of building and deploying models, translating raw data into actionable insights. • Skilled in identifying data science applications, cross-functional collaboration, mentoring team members. • Published 3 journal and multiple conference publications (85+ total citations).

Experience

  • Data Scientist 2 at Amazon
    Jul 2025 - Present · 1 yr 1 mo

  • Senior Data Scientist at Studio Sirah
    Jul 2021 - Jun 2025 · 4 yrs

    • Led the AI, Analytics & Data Engineering division; mentored four team members; collaborated with design, product & marketing teams. • Fine-tuned LLMs for game AI, built an embedding-based deck-builder, deployed a RAG pipeline. • Built SKU recommendation engine and trained intelligent bots using Deep Reinforcement Learning. • Setup cost-effective scalable data infrastructure on AWS, end-to-end analytics & dashboarding. • Ran large scale A/B experiments and supervised the marketing analytics

  • Data Science Intern at Pitney Bowes
    Jan 2021 - Jun 2021 · 6 mos

    Conducted advanced scenario analysis to identify network bottlenecks and optimize delivery times across the company’s complex logistics infrastructure.

  • Project Scientific Officer (Data Scientist Role) at Homi Bhabha Centre for Science Education (HBCSE), TIFR
    Aug 2016 - Jul 2019 · 3 yrs

    Functioned as a data scientist within the Learning Sciences Research Group, applying machine learning, statistical modeling, and neural network research to real-world behavioral and cognitive data. Projects combined analytics, simulation, and algorithm development to solve complex problems at the intersection of AI and human learning.

  • Developer Intern at Indiana University Bloomington
    May 2015 - Jul 2015 · 3 mos

    Interaction design • Embodied Learning • Graspable Math