Komal Diwe profile photo 1 on Tal

Komal Diwe

Data Scientist @ Tredence ● Generative AI & RAG System ● AWS

India

2+ years of work experience

Tredence Inc.
Jan 20269 mosPresent

Pune District

● Own the Conversation Management and NRT Personalization modules of a production RAG-based AI assistant for a leading US health & wellness brand — covering two-tier memory (Valkey → DynamoDB), LLM-as-a-judge evaluation, and a real-time CDC pipeline (DMS → Kinesis → Lambda). ● Contributed to a hybrid retrieval layer (semantic + keyword search, RRF fusion) grounding the assistant in the client's product catalog.
Tredence Inc.
Jul 20241 yr 6 mosDec 2025

Pune

Driving enterprise value by building end-to-end GenAI evaluation frameworks and statistical pricing models. Specializing in bridging the gap between raw model outputs and actionable business insights. 𝐆𝐞𝐧𝐀𝐈 & 𝐋𝐋𝐌 𝐄𝐧𝐠𝐢𝐧𝐞𝐞𝐫𝐢𝐧𝐠 * Designed a robust LLM evaluation framework to measure accuracy and relevance, successfully standardizing quality metrics for client deliverables. * Built and deployed AWS Lambda-based pipelines for Near Real-Time (NRT) data processing, and minimizing latency. * Created automated benchmarking tools and monitoring dashboards, enabling transparent model selection and performance tracking for business stakeholders. 𝐒𝐭𝐚𝐭𝐢𝐬𝐭𝐢𝐜𝐚𝐥 𝐌𝐨𝐝𝐞𝐥𝐢𝐧𝐠 & 𝐀𝐧𝐚𝐥𝐲𝐭𝐢𝐜𝐬 * Developed log-log regression models to analyze price elasticity, empowering retail clients to optimize pricing strategies based on statistical evidence. * Performed feature engineering on large-scale datasets (multi-million rows), ensuring high-fidelity input for pricing algorithms.
Argonaut AI
Apr 20239 mosDec 2023

San Francisco, CA

Responsibilities: Data Extraction, Data Labelling and analysis Used Pandas, Camelot, Python, Document AI, Google Cloud Platform Work: ● Improved LLM extraction accuracy by 20% through prompt and pipeline iteration on complex, unstructured documents. ● Engineered automated data-extraction and labeling pipelines using Document AI on GCP, converting messy source documents into structured, analysis-ready data. Other: ● Collaborating with a group comprising LLM Specialists, Machine Learning Engineers, and Data Scientists. ● Adapted to a fast-paced startup, efficiently and independently handling tasks beyond my assigned duties. ● 𝗣𝗹𝗮𝘆𝗲𝗱 𝗮 𝗸𝗲𝘆 𝗿𝗼𝗹𝗲 𝗶𝗻 𝘀𝗲𝘃𝗲𝗿𝗮𝗹 𝗽𝗿𝗼𝗷𝗲𝗰𝘁𝘀 𝘁𝗵𝗮𝘁 𝗰𝗼𝗻𝘁𝗿𝗶𝗯𝘂𝘁𝗲𝗱 𝘁𝗼 𝘁𝗵𝗲 𝗰𝗼𝗺𝗽𝗮𝗻𝘆'𝘀 𝘀𝘂𝗰𝗰𝗲𝘀𝘀𝗳𝘂𝗹 𝗮𝗰𝗾𝘂𝗶𝘀𝗶𝘁𝗶𝗼𝗻 𝗯𝘆 𝗚𝗼𝗼𝗴𝗹𝗲.

Komal’s AI Stack

claude
claude
hugging face
hugging face
langchain
langchain
langgraph
langgraph
openai
openai
claudecode
claudecode

Education

University of Hyderabad

Master of Technology - MTech, Artificial Intelligence

University of Hyderabad

20202 yrs2022

Bachelor of Technology, Information Technology

Shri Guru Gobind Singhji Institute of Engineering and Technology

20144 yrs2018

Join Komal and 10k+ other professionals on

Tal Boss

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

Download Tal Boss

Komal Diwe on Tal

Headline

Data Scientist @ Tredence ● Generative AI & RAG System ● AWS

About

𝐌𝐋 𝐄𝐧𝐠𝐢𝐧𝐞𝐞𝐫 & 𝐃𝐚𝐭𝐚 𝐒𝐜𝐢𝐞𝐧𝐭𝐢𝐬𝐭 | 𝐌.𝐓𝐞𝐜𝐡 𝐢𝐧 𝐀𝐈 | 𝐒𝐩𝐞𝐜𝐢𝐚𝐥𝐢𝐳𝐢𝐧𝐠 𝐢𝐧 𝐆𝐞𝐧𝐀𝐈, 𝐌𝐋𝐎𝐩𝐬 & 𝐏𝐫𝐨𝐝𝐮𝐜𝐭𝐢𝐨𝐧 𝐒𝐲𝐬𝐭𝐞𝐦𝐬 I'm a Data Scientist who 𝐛𝐫𝐢𝐝𝐠𝐞𝐬 𝐞𝐱𝐩𝐞𝐫𝐢𝐦𝐞𝐧𝐭𝐚𝐥 𝐀𝐈 𝐚𝐧𝐝 𝐬𝐜𝐚𝐥𝐚𝐛𝐥𝐞 𝐩𝐫𝐨𝐝𝐮𝐜𝐭𝐢𝐨𝐧 𝐬𝐲𝐬𝐭𝐞𝐦𝐬. With a background spanning a YC-backed startup and enterprise consulting, I build AI that doesn't just "work" in a notebook — it drives measurable business impact in production. Currently at Tredence, I work on a customer-facing, RAG-based AI assistant for a leading US health & wellness brand, where I own the modules that make it reliable and personalized: 𝐚𝐧 𝐋𝐋𝐌-𝐚𝐬-𝐚-𝐣𝐮𝐝𝐠𝐞 𝐞𝐯𝐚𝐥𝐮𝐚𝐭𝐢𝐨𝐧 𝐟𝐫𝐚𝐦𝐞𝐰𝐨𝐫𝐤 (𝐡𝐚𝐥𝐥𝐮𝐜𝐢𝐧𝐚𝐭𝐢𝐨𝐧/𝐪𝐮𝐚𝐥𝐢𝐭𝐲/𝐜𝐨𝐦𝐩𝐥𝐢𝐚𝐧𝐜𝐞) 𝐚𝐧𝐝 𝐚 𝐧𝐞𝐚𝐫-𝐫𝐞𝐚𝐥-𝐭𝐢𝐦𝐞 𝐩𝐞𝐫𝐬𝐨𝐧𝐚𝐥𝐢𝐳𝐚𝐭𝐢𝐨𝐧 𝐩𝐢𝐩𝐞𝐥𝐢𝐧𝐞 𝐨𝐧 𝐀𝐖𝐒 that keeps user context fresh within seconds. I also build statistical pricing models for retail clients. Previously, at Argonaut AI (YC-backed), I lived the zero-to-one journey — improving LLM extraction accuracy by 20% and engineering automated pipelines that turned complex unstructured documents into structured data. 🚀 Startup Agility & Enterprise Rigor — "ship fast" with enterprise scalability standards. 🛠️ Full-Stack ML — from a Jupyter notebook to a Dockerized, monitored production service. 🧰 GenAI Native — deep, hands-on RAG, vector search, and LLM evaluation. 📢 𝐎𝐩𝐞𝐧 𝐭𝐨 𝐌𝐋 𝐄𝐧𝐠𝐢𝐧𝐞𝐞𝐫, 𝐀𝐈 𝐄𝐧𝐠𝐢𝐧𝐞𝐞𝐫, 𝐚𝐧𝐝 𝐃𝐚𝐭𝐚 𝐒𝐜𝐢𝐞𝐧𝐭𝐢𝐬𝐭 𝐫𝐨𝐥𝐞𝐬 𝐟𝐨𝐜𝐮𝐬𝐞𝐝 𝐨𝐧 𝐆𝐞𝐧𝐀𝐈 𝐚𝐧𝐝 𝐩𝐫𝐨𝐝𝐮𝐜𝐭𝐢𝐨𝐧 𝐌𝐋 𝐬𝐲𝐬𝐭𝐞𝐦𝐬. Toolbox: Python · FastAPI · AWS (Lambda, Kinesis, DynamoDB, ECS, OpenSearch, Bedrock) · LLMs & RAG · Vector Search ·

Location

India

AI stack

claude

hugging face

langchain

langgraph

openai

claudecode

Work experience

Data Scientist at Tredence Inc., Jan 2026 - Present · 9 mos, Pune District, ● Own the Conversation Management and NRT Personalization modules of a production RAG-based AI assistant for a leading US health & wellness brand — covering two-tier memory (Valkey → DynamoDB), LLM-as-a-judge evaluation, and a real-time CDC pipeline (DMS → Kinesis → Lambda). ● Contributed to a hybrid retrieval layer (semantic + keyword search, RRF fusion) grounding the assistant in the client's product catalog.

Associate Data Scientist at Tredence Inc., Jul 2024 - Dec 2025 · 1 yr 6 mos, Pune, Driving enterprise value by building end-to-end GenAI evaluation frameworks and statistical pricing models. Specializing in bridging the gap between raw model outputs and actionable business insights. 𝐆𝐞𝐧𝐀𝐈 & 𝐋𝐋𝐌 𝐄𝐧𝐠𝐢𝐧𝐞𝐞𝐫𝐢𝐧𝐠 * Designed a robust LLM evaluation framework to measure accuracy and relevance, successfully standardizing quality metrics for client deliverables. * Built and deployed AWS Lambda-based pipelines for Near Real-Time (NRT) data processing, and minimizing latency. * Created automated benchmarking tools and monitoring dashboards, enabling transparent model selection and performance tracking for business stakeholders. 𝐒𝐭𝐚𝐭𝐢𝐬𝐭𝐢𝐜𝐚𝐥 𝐌𝐨𝐝𝐞𝐥𝐢𝐧𝐠 & 𝐀𝐧𝐚𝐥𝐲𝐭𝐢𝐜𝐬 * Developed log-log regression models to analyze price elasticity, empowering retail clients to optimize pricing strategies based on statistical evidence. * Performed feature engineering on large-scale datasets (multi-million rows), ensuring high-fidelity input for pricing algorithms.

Machine Learning Engineer at Argonaut AI, Apr 2023 - Dec 2023 · 9 mos, San Francisco, CA, Responsibilities: Data Extraction, Data Labelling and analysis Used Pandas, Camelot, Python, Document AI, Google Cloud Platform Work: ● Improved LLM extraction accuracy by 20% through prompt and pipeline iteration on complex, unstructured documents. ● Engineered automated data-extraction and labeling pipelines using Document AI on GCP, converting messy source documents into structured, analysis-ready data. Other: ● Collaborating with a group comprising LLM Specialists, Machine Learning Engineers, and Data Scientists. ● Adapted to a fast-paced startup, efficiently and independently handling tasks beyond my assigned duties. ● 𝗣𝗹𝗮𝘆𝗲𝗱 𝗮 𝗸𝗲𝘆 𝗿𝗼𝗹𝗲 𝗶𝗻 𝘀𝗲𝘃𝗲𝗿𝗮𝗹 𝗽𝗿𝗼𝗷𝗲𝗰𝘁𝘀 𝘁𝗵𝗮𝘁 𝗰𝗼𝗻𝘁𝗿𝗶𝗯𝘂𝘁𝗲𝗱 𝘁𝗼 𝘁𝗵𝗲 𝗰𝗼𝗺𝗽𝗮𝗻𝘆'𝘀 𝘀𝘂𝗰𝗰𝗲𝘀𝘀𝗳𝘂𝗹 𝗮𝗰𝗾𝘂𝗶𝘀𝗶𝘁𝗶𝗼𝗻 𝗯𝘆 𝗚𝗼𝗼𝗴𝗹𝗲.

Education

Master of Technology - MTech, Artificial Intelligence, University of Hyderabad

Bachelor of Technology, Information Technology, Shri Guru Gobind Singhji Institute of Engineering and Technology