Ayesha Siddiqa profile photo 1 on Tal

Ayesha Siddiqa

Applied AI @ Flipkart | Lead Data Scientist | Agentic systems · LLM fine-tuning & post-training · Multilingual NLP

Hyderabad

8+ years of work experience

Flipkart
Mar 20266 mosPresent

Bengaluru, Karnataka, India

Flipkart
Aug 20232 yrs 6 mosJan 2026

Bengaluru, Karnataka, India

Flipkart
Sep 20212 yrsAug 2023

Flipkart
Jul 20183 yrs 2 mosAug 2021

Bengaluru, Karnataka, India

IIT MADRAS
Jan 20143 yrs 1 mo2017

Greater Chennai Area

Ayesha’s AI Stack

antigravity
antigravity
codex
codex

Education

Indian Institute of Technology, Madras

Master’s Degree, Computer Science - Machine Learning and Natural Language Processing

Indian Institute of Technology, Madras

20143 yrs2017
Indian Institute of Technology, Madras

Master's degree, Computer Science

Indian Institute of Technology, Madras

Join Ayesha and 10k+ other professionals on

Tal Boss

Tal is the app where bosses meet engineers & builders directly, without middlemen

Download Tal Boss

Ayesha Siddiqa on Tal

Headline

Applied AI @ Flipkart | Lead Data Scientist | Agentic systems · LLM fine-tuning & post-training · Multilingual NLP

About

For ~8 years at Flipkart, I've worked at the intersection of applied research and production ML: conversational AI, LLMs and SLMs, RAG, Agentic AI systems, and multilingual NLP. I lead the CX Data Science team, where we own the AI behind post-order customer experience — from the chatbot that picks up your order context, to the copilot a human agent uses while talking to you. Some highlights from the journey so far: → The Agent Chat Copilot cut Average Handling Time by 50%, now runs 90K+ chats a day → Flipkart's in-house RAG framework (Instant Karma 2024) and a RAFT pipeline that beat GPT-3.5 on customer-support QA. → NER for Voice & Text Search / Decision Assistant — reduced support cost by 10% and delivered bottom-line savings. → Expanded vernacular NLP coverage from 5 to 12 Indian languages for agent-notes Summarization through Google collaboration utilising Gemini models for STT. Currently, my focus is agentic AI for post-purchase CX — moving from hardcoded orchestration to a reinforcement-learning-based system (DQN inside an agentic layer, treating the journey as an MDP) that learns from actual user feedback instead of static rules. Outside of shipping, I mentor junior scientists, translate research for business stakeholders (including the less glamorous but essential work of GenAI infra-cost estimation and ROI modelling), and publish — most recently accepted at SIGIR 2026, previously at KDD e-commerce 2022, Springer LNAI, and CICLING. Stack: PyTorch, HuggingFace, Haystack, Google ADK, Unsloth. Models I've fine-tuned and deployed in production: Llama, Mistral, Gemma 3, GPT/Gemini variants, BLIP2, ColQwen2, mBERT/mBART/XLM.

Location

Hyderabad

AI stack

antigravity

codex

Work experience

Senior Data Scientist 2 at Flipkart, Mar 2026 - Present · 6 mos, Bengaluru, Karnataka, India

Senior Data Scientist at Flipkart, Aug 2023 - Jan 2026 · 2 yrs 6 mos, Bengaluru, Karnataka, India

Data Scientist 3 at Flipkart, Sep 2021 - Aug 2023 · 2 yrs

Data Scientist 2 at Flipkart, Jul 2018 - Aug 2021 · 3 yrs 2 mos, Bengaluru, Karnataka, India

Research Scholar at IIT MADRAS, Jan 2014 - 2017 · 3 yrs 1 mo, Greater Chennai Area

Education

Master’s Degree, Computer Science - Machine Learning and Natural Language Processing, Indian Institute of Technology, Madras

Master's degree, Computer Science, Indian Institute of Technology, Madras

Ayesha Siddiqa - Senior Data Scientist 2 in Hyderabad | Tal