Kulothungan Senthil
Verified Expert in Engineering
AI Engineer and Developer
Chennai, Tamil Nadu, India
Toptal member since August 14, 2026
Kulothungan specializes in designing enterprise AI solutions that bridge business strategy with modern machine learning. With eight years of experience, he has built production-scale applications using Azure OpenAI, LangGraph, Databricks, and Azure AI Search. From multi-agent systems and contract intelligence to computer vision and forecasting, Kulothungan focuses on building reliable, explainable, and scalable AI products that deliver measurable outcomes.
Portfolio
Experience
- Data Scientist - 5 years
- Supervised Learning - 4 years
- Python - 4 years
- Agentic AI - 4 years
- Unsupervised Learning - 4 years
- AI Engineering - 4 years
- Azure DevOps - 4 years
- AI Agents - 4 years
Preferred Environment
Azure, Databricks, Azure DevOps, Agentic AI, Microsoft Power BI, Computer Vision, Predictive Analytics, AI Engineering, Data Scientist
The most amazing...
...solution I've developed is an AI contract intelligence platform that saved millions by detecting spend leakage.
Work Experience
Senior AI & ML Engineer
Mu Sigma Business Solutions
- Led the design and implementation of an enterprise-grade contract intelligence system, leveraging Azure Document Intelligence for layout-aware data extraction of tables, sections, and metadata.
- Spearheaded the development of a hybrid BM25 and HNSW retrieval architecture via Azure AI Search and LangGraph, utilizing advanced hierarchical chunking strategies to significantly enhance retrieval precision.
- Built agentic workflows with LangGraph to enable multi-step reasoning and explainable AI outputs, while establishing a rigorous evaluation framework using Recall@K, MRR, and RAGAS metrics.
- Balanced hands-on architecture and large language model (LLM) orchestration with the leadership of a six-engineer team, ultimately delivering multimillion-dollar savings through automated contract anomaly detection.
- Architected and executed a seamless end-to-end data engineering pipeline integrating Azure DevOps, Azure Data Lake, and Power BI to solve complex historical version tracking and lifecycle state challenges.
- Engineered analytical models to monitor core Agile delivery metrics, including cycle time, lead time, throughput, and velocity, resulting in the delivery of a top-25 most-utilized enterprise dashboard (Power BI).
- Built a RAG-based enterprise chatbot for real-time delivery intelligence and designed an embedding-based similarity engine.
- Utilized semantic vector space to automate duplicate work identification, reusability detection, and business capability clustering, significantly reducing redundant development efforts and optimizing resource allocation across the organization.
Data Scientist
Mu Sigma Business Solutions
- Architected and deployed high-performance computer vision pipelines for automated defect detection, utilizing YOLOv8 for real-time object detection and Detectron2 for precise instance segmentation.
- Engineered a robust hybrid training methodology by synthesizing Blender-generated imagery to simulate rare failure modes alongside real-world annotated datasets, ensuring model resilience through sophisticated preprocessing and augmentation.
- Defined rigorous annotation taxonomies and leveraged transfer learning to optimize models for both high mAP and inference efficiency, conducting deep-dive failure mode analysis and precision-recall evaluations to eliminate misclassifications.
- Enabled predictive maintenance workflows by optimizing models for superior generalization in diverse field conditions, slashing manual inspection overhead and enhancing operational uptime.
- Achieved the Chevron Human Energy Award for architecting and deploying a sophisticated time series and trend analysis suite across enterprise portfolio platforms, significantly enhancing visibility into project performance and cost evolution.
- Engineered a predictive analytics layer utilizing ARIMA and Random Forest Regressors to forecast project delays and cost overruns by identifying early-warning signals through deep feature engineering on phase-wise velocity and historical deviation.
- Implemented Monte Carlo simulations for what-if scenario planning, enabling stakeholders to optimize resource allocation and balance demand versus capacity with high precision.
- Owned the end-to-end lifecycle from requirement gathering to production deployment, translating complex forecasting outputs into actionable planning decisions that drove measurable operational efficiency across the global portfolio.
Machine Learning & IoT Embedded Engineer
Smartx Connected Products
- Built embedded IoT solutions for products, including pet trackers, smart food machines, and medical storage systems using STM, NUCLEO, and BLE modules by integrating sensors, communication modules, and embedded controllers for real-time monitoring.
- Performed data scraping, preprocessing, and analytics using Python, Excel, and Power BI to support firmware interfacing, telemetry analysis, and operational reporting.
- Developed edge machine learning (ML) models for edge IoT devices for real-time prediction.
Junior Embedded Engineer
Inspire Clean Energy
- Contributed to the development of a solar soiling analysis robot (SABOT) through mechanical, electronics, and embedded systems integration for solar panel monitoring.
- Designed data collection pipelines and visualization dashboards using Excel and Tableau for operational and performance analysis.
- Implemented remote connectivity, sensor integration, and embedded telemetry systems to support real-time monitoring and analytics workflows.
Experience
Face Detection and Tracking
Education
Bachelor's Degree in Mechatronics
Shanmugha Arts, Science, Technology, and Research Academy (SASTRA) - Thanjavur, India
Certifications
SQL Server Integration Services
Dataiku Core Designer
Dataiku
Skills
Libraries/APIs
PyTorch, NumPy, TensorFlow, XGBoost, ExcelJS, Hugging Face Transformers, REST APIs
Tools
Microsoft Power BI, Blender, Tableau, Azure OpenAI Service, Power Query
Languages
Python, SQL, C++, Embedded C++
Paradigms
Azure DevOps
Frameworks
LightGBM
Platforms
STM32, Azure, Databricks, Docker
Other
Supervised Learning, Unsupervised Learning, Agentic AI, AI Engineering, Data Scientist, Azure Data Lake, YOLOv8, Detectron2, Deep Learning, Model Evaluation, Large Language Models (LLMs), Instance Segmentation, Convolutional Neural Networks (CNNs), Random Forests, Feature Engineering, AI Agents, AutoARIMA, Prompt Engineering, LangChain, Retrieval-augmented Generation (RAG), FAISS, DeepEval, Responsible AI, MLflow, FastAPI, Model Serving, CI/CD Pipelines, DAX, Data Modeling, Computer Vision, Predictive Analytics, Mechatronics, Electronics, Data Science, Data Engineering, Full-stack, Data Analysis, Unsupervised Machine Learning, Clustering, Dashboards, Portfolio Management, Web Applications
How to Work with Toptal
Toptal matches you directly with global industry experts from our network in hours—not weeks or months.
Share your needs
Choose your talent
Start your risk-free talent trial
Top talent is in high demand.
Start hiring