IPSG Systems has expertise in helping clients develop solutions based on machine learning algorithms as well as in several open source Big Data frameworks like Spark and Hadoop. We have developed several cutting edge solutions by deploying algorithms and technologies like Deep Neural networks for image recognition, compute-intensive methods like L-BFGS optimization for retail assortment planning and so on. Our team comprises of gifted, highly talented engineers who can provide a machine learning algorithm that can work on single machine as well as in multi-node clusters that run on Spark and Hadoop.
Our expertise stems not only from a deep understanding of these frameworks, but from a strong foundation in machine learning concepts, data science, and distributed programming. Several members in our team are actively involved in contributing to open source projects in various domains that include compilers (LLVM), distributed programming and Spark
Algorithms and Technologies
Deep Networks
- Recursive Neural networks
- Convolution networks
Classification / Regression
- Support Vector Machines
- Decision Trees
- Neural Networks
Optimization
- Gradient descent
- BFGS / L-BFGS
- Hill Climbing, Simulated Annealing
Industry
Retail
- Assortment planning, demand forecasting
Insurance
- Risk analysis
Data centre analytics
- Root cause analysis, time series data analysis
IoT
- Anomaly and outlier detection
IT Ticketing
- Applications in clinical research