CONVERGING_
000Iteration
—Loss
0.100Learn Rate
INITPhase
0%
S

ROHIT
SAROJ

Sr. Data Scientist — Generative AI · LangGraph · AWS · RAG · LLMs

Building AI systems that ship to production
and solve real business problems.

Let's Talk↗
Scroll
LangGraphAWS BedrockGenerative AI RAG PipelinesDSPyLangChain Deep LearningFastAPISnowflakePySpark LangGraphAWS BedrockGenerative AI RAG PipelinesDSPyLangChain Deep LearningFastAPISnowflakePySpark
OpenAITensorFlowPyTorch PineconeScikit-learnPower BI DockerJenkins CI/CDPandasXGBoost OpenAITensorFlowPyTorch PineconeScikit-learnPower BI DockerJenkins CI/CDPandasXGBoost
001 — Profile
AI
THAT
SHIPS
TO PROD
8+
Years Exp.
5
Companies
40%
Res. Time ↓
1.5K
Medium Followers

Experienced Data Scientist with a track record across energy, cleantech, and healthcare AI. I specialise in Generative AI systems — building agents that work in production, not just notebooks.

My stack: LangGraph, DSPy, AWS Bedrock, LangChain, OpenAI — backed by deep classical ML, forecasting, and data engineering built over 8 years.

Currently designing a clinical cohort-building AI agent at Teamlease. Previously cut issue resolution time by 40% at Stem Inc with a RAG system.

WORK
HISTORY

05 Roles · 2017 → Present
01 / 05
SR. DATA SCIENTIST
Teamlease
Aug 2025 – Present · Gurgaon, IN
GenAI-powered RWE Agent via LangGraph for clinical cohort building. Prompt optimisation with DSPy, Claude 3.5 via VOX API. AWS KB + Snowflake schema mapping. Human-in-the-loop review workflows.
LangGraphDSPyAWS BedrockSnowflakeFastAPI
02 / 05
DATA SCIENTIST
Stem Inc
Jun 2023 – Apr 2025 · Gurgaon, IN
RAG system using OpenAI embeddings + Pinecone — resolution time down 40%. Enhanced site load forecasting, automated analytics pipelines (25% efficiency gain), AWS + Jenkins CI/CD.
LangChainOpenAI APIPineconeJenkinsAWS
03 / 05
ANALYST
Manikaran Analytics
Sep 2022 – Jun 2023 · New Delhi, IN
Refined wind energy forecasting — MAE down 12%. Wind speed prediction with XGBoost & Random Forest. Automated tasks saving 25% team resources.
XGBoostRandom ForestPandasSQL
04 / 05
PERFORMANCE ANALYST
Emergya Wind Turbines
Mar 2020 – Aug 2022 · Chennai, IN
CNN model for turbine power curve classification — 50% efficiency improvement. ML-based failure prediction, 20% maintenance cost reduction. Python GUI automation.
TensorFlowCNNPySparkPower BI
05 / 05
ENGINEER
Wind World India Ltd
Mar 2017 – Mar 2020 · Mumbai, IN
5% energy yield improvement through turbine performance analysis and root cause investigation. Foundation in data-driven engineering practice.
ExcelSQLPython

SKILLS

Technical competencies
across the full ML stack
ExpertinPython, buildingGenerativeAI agentswithLangGraph andDSPy,deploying RAGpipelineson AWSBedrock,training deeplearningmodels withTensorFlowandPyTorch.
Programming
Python
SQL
Git / Docker
Jenkins / CI/CD
Bash / Linux
Generative AI
LangGraph
LangChain
DSPy
OpenAI / Bedrock
RAG / Pinecone
ML / Data
TensorFlow / PyTorch
Scikit-learn / XGBoost
Pandas / NumPy
PySpark
EDA / Feature Eng.
Cloud / Infra
AWS S3 / EC2
AWS Bedrock / KB
Snowflake
FastAPI / Streamlit
Power BI / Plotly
EDUCATION
2022
PGDBM — Operations Management
NMIMS, Mumbai
2016
B.E. — Instrumentation Engineering
Rajiv Gandhi Institute of Technology, Mumbai
CERTIFICATIONS
01
Prompt Design in Vertex AI
Google
02
SQL for Data Science
Great Learning
03
The Data Science Course 2022
Udemy
04
Generative AI with LangChain
Udemy
005 — Get in Touch
LET'S WORK TOGETHER
Direct Links
rohitsaroj29@gmail.com↗ LinkedIn — rohitsaroj↗ Medium — 1.5K+ Followers↗
© 2025 Rohit Saroj
All rights reserved
Gurgaon, India
Vanya Avatar

Vanya

Online
Hey there! 👋 I'm Vanya, Rohit's AI assistant. Ask me anything about his experience, skills, or how to connect with him!
```