Trend Pilot
Multi-agent assistant for creator teams that catches trends, generates scripts, and assigns work. Built with FastAPI, Next.js, GPT-4o VLM, and CI/CD.
- 25→60%
- post hit rate
- ~22×
- modeled ROI
Open to full-time roles · Los Angeles, CA
Shamik Basu
AI Enabled Data Scientist
I ship production ML, RAG, and LLM systems with measurable impact: 72% lower cost on a 5M+ record/month pipeline, 27% better fraud detection, and 115× GPU speedups. MS Applied Data Science at USC (GPA 3.7).
- 2+ years production ML
- MS Data Science · USC
- 1st Place · UCLA SAIRS
Highlights
Projects
Featured Projects
End-to-end builds with case studies, demos, and measurable outcomes.
1st Place
Hackathon · 24 hrs
EcoMate-AI
Multimodal agentic platform (Gemini 2.5 Flash) that turns receipts and activity into sustainability actions.
- UCLA SAIRS
- 2025 winner
115× faster
GPU Computing
Custom CUDA Library
Shared-memory tiled kernels with Python bindings, delivering 115× CPU and 14× PyTorch speedups on Tesla T4.
Data Viz
Workflow Analytics
Smooth Operation
Atlassian Forge app with heatmaps, analytics, smart search, and vulnerability scanning for Jira and Confluence.
Experience
Work Experience
Impact-first summaries. Full detail in the resume PDF.
-
Data Science Associate Intern
Liquidity AI Capital · Los Angeles, CA
Deployed production SLM inference in Docker with end-to-end LoRA fine-tuning for domain-specific financial inference.
-
Data Science Associate Intern
KCC Capital Partners · Los Angeles, CA
Built an agentic LangChain pipeline for M&A target sourcing with LLM classification, web scraping, and structured extraction.
-
Data Scientist
Bajaj Finserv Health · Pune, India
- 72% cost reduction
- 5M+ records/mo
- 92% accuracy
Production RAG pipeline (LangChain + GPT-3.5) with Langfuse observability and Kafka event-driven inference.
-
Associate Data Scientist
Bajaj Finserv Health · Pune, India
- +27% fraud accuracy
- +22% efficiency
Redesigned a production NER pipeline and shipped supervised models with fewer false positives.
-
Data Engineer Intern
Bajaj Finserv Health · Pune, India
Scaled ETL to 10M+ records across 200+ features with C++/Trino/Docker and optimized SQL.
Education
Education
Applied data science and computer science foundations.
University of Southern California
MS, Applied Data Science
- Deep Learning
- Machine Learning
- Data Mining
- Data Management
SRM Institute of Science & Technology
BTech, Computer Science
- Algorithms
- Artificial Intelligence
- Computer Vision
- Probability & Queueing
Leadership
Leadership
Community building and graduate mentorship at USC.
250+
Vice President
GRIDS, Graduate Data Science Club, USC
Lead analytics workshops, ideathons, and projects for a 250-member graduate data science community.
1:1
Graduate Mentor
USC Viterbi School of Engineering
Mentor graduate students on data science career paths, technical communication, and professional development.
Skills
Technical Skills
Core tools and methods I use in production work.
Python, SQL, R, C/C++, PostgreSQL, MongoDB, Spark, Trino, Kafka, Machine Learning, Statistics, Causal Inference, A/B Testing, NLP, NER, Forecasting, Tableau, Power BI, PyTorch, TensorFlow, Hugging Face, LoRA, RAG, LangChain, Multi-Agent Systems, AWS, Docker, Kubernetes, REST APIs, CI/CD, Azure Synapse, Langfuse, Git
Languages & Databases
- Python
- SQL
- R
- C/C++
- PostgreSQL
- MongoDB
- Spark
- Trino
- Kafka
ML & Analytics
- Machine Learning
- Statistics
- Causal Inference
- A/B Testing
- NLP
- NER
- Forecasting
- Tableau
- Power BI
Deep Learning & LLMs
- PyTorch
- TensorFlow
- Hugging Face
- LoRA
- RAG
- LangChain
- Multi-Agent
Cloud & DevOps
- AWS
- Docker
- Kubernetes
- REST APIs
- CI/CD
- Azure Synapse
- Langfuse
- Git
Contact
Contact
Email is best. I typically reply within 24 hours.