Staff Applied Scientist & AI Architect

Sarthak
Mishra

Building production-grade Generative AI systems — multi-agent orchestration, LLM inference, and enterprise RAG pipelines.

sarthak405@gmail.com +91 95609 22269 linkedin.com/in/sarthak405
IIT Delhi · CS & Engineering
AI
01

Experience

Work History
Staff Applied Scientist
Tekion Corp
Feb 2026 – Present
  • Leading AI/ML initiatives within Tekion's automotive retail technology platform.
  • Designing and building intelligent data science systems for dealership operations and customer experience.
GenAI Automotive Tech AI Agents LangGraph
Advisory Data Scientist
IBM
Jul 2024 – Jan 2026
  • Led and mentored a team of 6–8 data scientists and ML engineers, owning sprint planning, technical direction, and delivery for multiple Generative AI initiatives.
  • Spearheaded a commercialized SAP GenAI product automating Technical Specification generation using GPT-4o, multimodal document understanding, and custom RAG pipelines — 30–40% productivity improvement.
  • Designed multi-agent GenAI systems using LangChain, LangGraph, AutoGen, and Strands, deployed on AWS and Azure.
  • Guided team on LLM fine-tuning (PEFT/LoRA), quantization (4-bit), and RoPE scaling for lower inference costs.
  • Led deployment of LLM inference endpoints on AWS EC2, integrating Amazon Bedrock, Azure OpenAI, and IBM watsonx.ai.
LangGraph LangChain AutoGen GPT-4o AWS Azure RAG PEFT / LoRA Bedrock watsonx.ai
Senior Data Scientist
IBM
Jul 2019 – Jun 2024
  • Designed end-to-end ML solutions across time-series forecasting, optimization, computer vision, and audio analytics.
  • Implemented transformer-based forecasting models (PatchTST, TimesFM) for sales and inventory planning; achieved MAPE of 0.19.
  • Solved linear and quadratic optimization problems using CPLEX, PuLP, and QP solvers for operational decision-making.
  • Implemented computer vision and audio ML pipelines (GoogLeNet, YOLOv5, GMM-based models) within IBM Maximo Visual Inspection, achieving ~90% accuracy.
PatchTST TimesFM YOLOv5 CPLEX Computer Vision Forecasting IBM Maximo
02

Publications & Recognition

PUB
Co-authored peer-reviewed research — "Navigating ethical minefields: a multi-stakeholder approach to assessing interconnected risks in Generative AI using grey DEMATEL" — published in Frontiers in Artificial Intelligence.
SPK
Invited speaker at GenAI ML Conclave – Bangalore (Feb 2025), presenting on agentic AI systems, Computer-Using Generalist Agents (CUGA), and emerging GenAI architectures.
SPK
Selected to represent Agentic AI and AI Governance at the IBM Consulting and IBM Technology Conclave, engaging with 200+ enterprise clients and leaders.
CERT
Certified Data Scientist – Level 1, The Open Group. Completed MIT xPRO Quantum Computing Fundamentals (4 CEUs). Completed Watsonx Orchestrate Technical Bootcamp in enterprise multi-agent orchestration.
WRK
Participated in the AISG and AIIS Asset and Offering Workshop, focused on scalable and ethical AI solution design.
03

Education

B.Tech — Computer Science & Engineering
Indian Institute of Technology, Delhi
2019