Looks like you have JavaScript disabled. For the full Twine experience, you will need to re-enable it.

AI/ML Engineer with 4+ years of experience building production-scale intelligent systems and full-stack applications. Proven expertise in classical ML, deep learning, LLMs, and agentic AI frameworks (LangChain, LangGraph, CrewAI) with strong MLOps capabilities. Successfully translated business requirements into scalable technical solutions at Adobe, delivering ML forecasting systems with 85% accuracy across 1M+ enterprise customers. Proficient in Python, PyTorch, ReactJS, AWS, Azure Databricks, Snowflake, and end-to-end model deployment. Master's in AI from NUS (Dean's Merit List). Seeking challenging roles in AI/ML engineering, MLOps, or full-stack AI development.…AI/ML Engineer with 4+ years of experience building production-scale intelligent systems and full-stack applications. Proven expertise in classical ML, deep learning, LLMs, and agentic AI frameworks (LangChain, LangGraph, CrewAI) with strong MLOps capabilities. Successfully translated business requirements into scalable technical solutions at Adobe, delivering ML forecasting systems with 85% accuracy across 1M+ enterprise customers. Proficient in Python, PyTorch, ReactJS, AWS, Azure Databricks, Snowflake, and end-to-end model deployment. Master's in AI from NUS (Dean's Merit List). Seeking challenging roles in AI/ML engineering, MLOps, or full-stack AI development.

Sneha Sarkar

Data Scientist, AI Engineer, Full Stack Developer, +4





AI/ML Engineer with 4+ years of experience building production-scale intelligent systems and full-stack applications. Proven expertise in classical ML, deep learning, LLMs, and agentic AI frameworks (LangChain, LangGraph, CrewAI) with strong MLOps capabilities. Successfully translated business requirements into scalable technical solutions at Adobe, delivering ML forecasting systems with 85% accuracy across 1M+ enterprise customers. Proficient in Python, PyTorch, ReactJS, AWS, Azure Databricks, Snowflake, and end-to-end model deployment. Master's in AI from NUS (Dean's Merit List). Seeking challenging roles in AI/ML engineering, MLOps, or full-stack AI development.…AI/ML Engineer with 4+ years of experience building production-scale intelligent systems and full-stack applications. Proven expertise in classical ML, deep learning, LLMs, and agentic AI frameworks (LangChain, LangGraph, CrewAI) with strong MLOps capabilities. Successfully translated business requirements into scalable technical solutions at Adobe, delivering ML forecasting systems with 85% accuracy across 1M+ enterprise customers. Proficient in Python, PyTorch, ReactJS, AWS, Azure Databricks, Snowflake, and end-to-end model deployment. Master's in AI from NUS (Dean's Merit List). Seeking challenging roles in AI/ML engineering, MLOps, or full-stack AI development.

Available to hire

AI/ML Engineer with 4+ years of experience building production-scale intelligent systems and full-stack applications. Proven expertise in classical ML, deep learning, LLMs, and agentic AI frameworks (LangChain, LangGraph, CrewAI) with strong MLOps capabilities. Successfully translated business requirements into scalable technical solutions at Adobe, delivering ML forecasting systems with 85% accuracy across 1M+ enterprise customers. Proficient in Python, PyTorch, ReactJS, AWS, Azure Databricks, Snowflake, and end-to-end model deployment. Master’s in AI from NUS (Dean’s Merit List). Seeking challenging roles in AI/ML engineering, MLOps, or full-stack AI development.

Skills

Experience Level

Expert

Expert

Expert

Expert

Expert

Expert

Expert

Expert

Expert

Expert

Expert

Expert

Expert

Expert

Intermediate

Intermediate

Language

English

Fluent

Work Experience

Research Assistant at Dean’s Office (Yale-NUS College)

May 1, 2025 - November 1, 2025

Engineered multi-stage OCR pipeline to digitize 38,500+ ship registry entries (423,500+ data fields) from a 770-page 19th-century maritime archive, achieving 80% automated extraction accuracy and reducing manual digitization time from ~1,987 hours to ~400 hours; developed computer vision preprocessing for degraded documents; implemented ensemble OCR (Google DocAI, Tesseract, PaddleOCR) with postprocessing for 11-column ship-level metadata extraction to enable searchable archives.

Teaching Assistant at National University of Singapore

May 1, 2025 - Present

Designed and led tutorials for CS3244 Machine Learning: Design, covering core ML concepts, algorithms, and practical workflows; mentored project teams. Used Python, NumPy, pandas, scikit-learn, and Matplotlib for data handling and model evaluation. Contributed to CeNCE AI Pathway labs: kNN, Decision Trees, Ensembles, Linear Models, and model evaluation; created Colab notebooks, exercises, and Kaggle-style competitions to reinforce hands-on learning.

Research Assistant at Dean's Office (Yale-NUS College)

May 1, 2025 - November 1, 2025

Engineered a multi-stage OCR pipeline to digitize 38,500+ ship registry entries (423,500 data fields) from a 770-page 19th-century maritime archive. Achieved 80% automated extraction accuracy and reduced manual digitization time from ~1,987 hours to ~400 hours, delivering ~5x efficiency gains. Implemented an ensemble OCR approach combining Google Document AI, Tesseract, and PaddleOCR with custom post-processing (error correction, confidence scoring, field validation) to extract 11-column structured ship metadata for a searchable digital archive. Addressed degraded document quality and handwriting variability for robust results.

SDE II - Enterprise Data Analytics at Adobe

August 1, 2020 - June 1, 2024

Led the development of a production ML forecasting system on Azure Databricks for customer churn prediction and product usage analytics. Achieved 85% prediction accuracy across 1M+ enterprise customers using ensemble models (Prophet, ARIMA, N-BEATS, Naive D rift) with dynamic per-customer hyperparameter tuning. Orchestrated end-to-end MLOps via Apache Airflow with Power BI dashboards enabling data-driven retention strategies (targeted discounts, engagement campaigns) to reduce churn. Architected secure data access using Microsoft Power Platform (Power Apps, Power Automate), enabling 30+ cross-functional teams to query 1M+ records.