Soham Dandapath

Applied AI Engineer at C3 AI

I turn messy, real-world business problems into AI systems that ship to production.

Redwood City, CA · MS @ Columbia

About

I'm an Applied AI Engineer at C3 AI and a relentless AI enthusiast. Most of my work lives where applied research meets solid engineering: I take ideas from problem framing all the way to production, mostly RAG-based LLM systems and probabilistic time-series forecasting on C3's agentic platform.

I care about models that are interpretable, deployments that are reproducible, and tooling that makes the next person's work easier. Here is where I focus:

RAG & LLM systemsLow-latency, policy-compliant retrieval and generation over enterprise data.
Time-series forecastingProbabilistic, hierarchical, explainable models that drive real decisions.
ML tooling & deploymentReproducible pipelines and toolchains that get models to production fast.
$0.8B
Est. annual impact, enterprise forecasting program
$7M+
Value delivered across customer forecasting apps
Hours → min
Faster deploys from internal tooling I built
30+
Open-source repositories on GitHub

Experience

2024 - Present
Applied AI Engineer · C3 AI

Lead applied ML and LLM projects end-to-end for global enterprises, currently an enterprise forecasting program for a major semiconductor customer with ~$0.8B estimated annual impact. Shipped a RAG document-retrieval system, demand and yield forecasting apps, and built an internal deployment toolchain that cut deploys from hours to minutes. I also mentor data scientists across teams.

2023
Data Science Intern · C3 AI

Built an out-of-the-box hierarchical forecasting and reconciliation system (MinT/ERM, DeepVAR-Hierarchical) and added Integrated Gradients explainability to probabilistic forecasts.

2022
Data Scientist · Charles & Keith

Tree-based sales forecasting, a 95%+ accuracy image-similarity engine for product matching, and an order-management app that cut costs and stockouts.

2020 - 2021
Earlier · Shopee, Seagate, Outstrip, CogniAble

Data pipelines on Airflow/HDFS, ML for hard-drive test-time prediction, KPI dashboards, and an I3D action-recognition model for early autism screening.

Selected Work

See all 30 repositories on GitHub ↗

Skills & Tools

Languages
Python · SQL · JavaScript
ML / AI
RAG & LLM systems · Time-series forecasting · Probabilistic ML · Distributed training · PyTorch
Infra / Tools
AWS / SageMaker · Airflow · CI/CD · Git

Education

2022 - 2023

MS, Computer Science · focus in Machine Learning

2018 - 2022

BE, Computer Science

Let's build something.

I'm always up for interesting problems and good conversations, whether that's a role, a collaboration, or just comparing notes on AI.

Say hello