Soham Pathak Email & Phone Number

Bengaluru, Ka, In

Bengaluru, Karnataka, India

Bengaluru, Karnataka, India

1. Led the development of a comprehensive End-to-End API using Custom LLM models for realtime log analysis .Tools and Technologies Used : Langchain , OpenAI , kubeCTL , Kubernaties , Seweed(AWS) , FastAPI , PG-Vector , PostgresDB2. Implementation of a streamlined Data Version Control (DVC) pipeline, optimizing the integration process for both existing and novel projects -- 50% reduction in time-to-liveTools and Technologies Used : DVC , MLFlow , FastAPI , MongoDB3. Real time incident correlation using incremental learning - Predict and give real time corelated incidents Tools and Technologies Used : DataSketch,MongoDB,Clustering Models(DBScan , KNN etc) , FastAPI4. Enhance the predictive capabilities of existing models by implementing fast data processing ,optimal encoders and industry-leading feature Engineering Techniques.Tools and Technology : SKLearn , DataSketch

Payment Date Prediction (Collection Use case):● Worked on 5+ live accounts data● Did analysis and created features to predict invoice DTP● Conceptualized RFM Metric to provide explainability to prediction accuracy and customer profiling.Rule engine for bank file processing (Treasury Usecase):● Built an engine to tag bank data with respective cashflow category● Standardized the bank rules recording process● Decreased the end to end processing time by 10xPayment Stoppage detection algorithm (Anomaly Detection):● Created a process to filter out customers who stopped paying to the client by using past-due amount buildup and Z score● Improved model accuracy by removing the outlier customersAuto rule inference:● Built a heuristic approach for classifying bank records into different cashflow categories using record description (machine-generated data)● Speed up the initial rule discovery processData Health Monitoring Using - Elasticsearch & Kibana (EPK):● Created a framework around the Elasticsearch package to pull data from the database Increased efficiency by removing the manual effort of querying and comparing data from the databaseCompute optimisation in AWS:● Help design pipeline for parallel multi-model training in AWS using SQS,lambda and EC2

India

Bhubaneshwar Area, India

The society devoted to bridging the gap between people's awareness and the world of AI.

Bhubaneshwar Area, India

Bhubaneshwar Area, India

Cloud Computing(Red Hat Open Stack Platform)

worked as a data analyst and ML Engg . Data pipeline and anomaly detection engine .

Bhubaneshwar Area, India

It's been two year since I started working with IoT and till date I have completed four projects regarding automation and surveillance .

New Delhi Area, India

Rapid web build (both front-end & back-end)