Shwet K Mishra profile photo 1 on Tal

Shwet K Mishra

AI/ML @Salesforce

Greater Bengaluru Area

10+ years of work experience

Salesforce
Apr 20251 yr 5 mosPresent

Bengaluru

Docket
Jul 20249 mosMar 2025

Bengaluru

- Built a real-time multi-agent AI sales copilot assisting Account Executives during live customer calls. - Designed orchestration pipelines combining intent detection, retrieval, and memory for dynamic task execution. - Developed a semantic cache with multi-vector retrieval, improving cache hit rate and reducing latency.
Freshworks
Apr 20231 yr 4 mosJul 2024

Bengaluru

- Led development of a production-grade RAG conversational agent for ITSM automation serving high-volume enterprise workflows. - Built a deterministic LLM evaluation framework combining automated grading and human-aligned scoring for safe production rollout. - Directed fine-tuning of open-source LLMs (7B–70B) using PEFT in PyTorch, reducing inference cost and improving deployment efficiency. - Partnered with infrastructure and product teams to deploy scalable GenAI systems across enterprise environments.
Freshworks
May 20212 yrsApr 2023

Bengaluru

- Built a streaming AIOps anomaly detection engine for large-scale time-series monitoring. - Designed an alert recommendation system, improving signal quality and reducing manual tuning. - Delivered a Smart Replies system, reducing response time in ITSM workflows.
DataWeave
Mar 20192 yrs 3 mosMay 2021

Bangalore

- Built solutions pertaining to Attribute-Value extraction and normalization from the e-commerce product title and descriptions - Built Retail Knowledge Graph to store dark unstructured data from the web to a structured graph DB(Neo4j) - Built an Encoder-Decoder based deep learning solution to categorize e-commerce products to their respective taxonomy
Exzeo
Aug 20162 yrs 8 mosMar 2019

Noida Area, India

- Built Risk Modelling system to classify risky and non-risky properties for home insurance domain - Built Image classification tool to predict roof shape out of Flat, Gable and Hip from their satellite images. - Planned and developed a solution for a project to extract meaningful information from a variety of insurance documents(.pdf,.txt, mails). - Built product for insurance underwriters to understand Claim Cost w.r.t. Claim type and Property features.
GitHub
@shwetkm

Shwet’s AI Stack

claude code
claude code
hugging face
hugging face
huggingface
huggingface
DSPy
DSPy
Pinecone
Pinecone
vLLM
vLLM

Education

International Institute of Information Technology Bangalore

MS by Research, Data Science

International Institute of Information Technology Bangalore

Join Shwet and 10k+ other professionals on

Tal Boss

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

Download Tal Boss

Shwet K Mishra on Tal

Headline

AI/ML @Salesforce

About

Driving research and applied development in Generative AI and Large Language Models (LLMs), with expertise in NLP, NLG, Time Series, and Multi‑Modal AI. Over 9 years of experience delivering AI solutions across SaaS, Enterprise, and Consumer products. Current focus on advancing enterprise‑ready agentic systems, multimodal reasoning, and trustworthy AI.

Location

Greater Bengaluru Area

AI stack

claude code

hugging face

huggingface

DSPy

Pinecone

vLLM

Work experience

Lead Member of Technical Staff - AI/ML at Salesforce, Apr 2025 - Present · 1 yr 5 mos, Bengaluru

Senior Machine Learning Engineer at Docket, Jul 2024 - Mar 2025 · 9 mos, Bengaluru, - Built a real-time multi-agent AI sales copilot assisting Account Executives during live customer calls. - Designed orchestration pipelines combining intent detection, retrieval, and memory for dynamic task execution. - Developed a semantic cache with multi-vector retrieval, improving cache hit rate and reducing latency.

Lead Data Scientist at Freshworks, Apr 2023 - Jul 2024 · 1 yr 4 mos, Bengaluru, - Led development of a production-grade RAG conversational agent for ITSM automation serving high-volume enterprise workflows. - Built a deterministic LLM evaluation framework combining automated grading and human-aligned scoring for safe production rollout. - Directed fine-tuning of open-source LLMs (7B–70B) using PEFT in PyTorch, reducing inference cost and improving deployment efficiency. - Partnered with infrastructure and product teams to deploy scalable GenAI systems across enterprise environments.

Senior Data Scientist at Freshworks, May 2021 - Apr 2023 · 2 yrs, Bengaluru, - Built a streaming AIOps anomaly detection engine for large-scale time-series monitoring. - Designed an alert recommendation system, improving signal quality and reducing manual tuning. - Delivered a Smart Replies system, reducing response time in ITSM workflows.

Data Scientist at DataWeave, Mar 2019 - May 2021 · 2 yrs 3 mos, Bangalore, - Built solutions pertaining to Attribute-Value extraction and normalization from the e-commerce product title and descriptions - Built Retail Knowledge Graph to store dark unstructured data from the web to a structured graph DB(Neo4j) - Built an Encoder-Decoder based deep learning solution to categorize e-commerce products to their respective taxonomy

Data Scientist at Exzeo, Aug 2016 - Mar 2019 · 2 yrs 8 mos, Noida Area, India, - Built Risk Modelling system to classify risky and non-risky properties for home insurance domain - Built Image classification tool to predict roof shape out of Flat, Gable and Hip from their satellite images. - Planned and developed a solution for a project to extract meaningful information from a variety of insurance documents(.pdf,.txt, mails). - Built product for insurance underwriters to understand Claim Cost w.r.t. Claim type and Property features.

Education

MS by Research, Data Science, International Institute of Information Technology Bangalore

Public links

github: https://github.com/shwetkm