Starting and Leading AI Charter @ Klub- Setting up the ML Platform for experimentation using dvc and MLFlow- Led a ML Solution for Auto Categorisation of Financial Statements using Custom NER Model, thereby helping in operational efficiency. - Built and Architected a Conversational Voice LLM with the capability to talk to human user helping in the lead generation. - Built a Telephony Voice LLM to act as an SDR agent for issue resolution.- Built a Retrieval-Augmented Generation (RAG) system utilizing a fine-tuned, quantized 7B model on MLX.

Leading Data and Data Infra Platform- Started the Data Charter in the company, worked as an IC to leading a team of 2 people. - Built up a Data Streaming solution using Self Hosted Kafka Connect which helped in data ingestion which eliminated the 3600$ / year.- Setting up Debezium for CDC to data replications which aided in data migrations and data storage. - Set up a GraphQL service which is used extensively in the company which has increased velocity of development team and have reduced the TAT for cross team collabration. - Building up Custom Kafka Connect SMTs according to our data usages and creation of kafka connect cli for generation of KF configs and management. - Migration of Airflow to K8s. - Setting up Data Quality rules over greatexpectations and transformation stage over dbt.- Building up a data lake and optimisation of revenue data storage which led to cost reduction of 12000$/ year.- Setting up Trino as Query Engine and HMS for efficient query capability.

Leading Loan Suite of ServiceThis involved four different services -> Loan, NACH, Repayments and Reconcilation.- Spearheaded a critical repayment service, ensuring real-time payouts to brands and vendors, crucial for financial operations which handles more than million transactions per month. (NR: Availability and Scalability)- Developed a reconciliation engine, reducing turnaround time (TAT) for payouts reconciliation from 5 days to just an hour, enhancing operational efficiency.- Led the ground-up development of a Loan Service, from schema design to service architecture- Successfully managed stakeholders across multiple verticals, establishing SLAs for service delivery.- Implemented architectural changes to accommodate a cloud-neutral to prevent vendor lock-in. - Optimized service performance by identifying and rectifying anti-patterns in database queries, resulting in reduced memory footprint.- Led a team of 5 people.

Led Revenue Pod - Responsible for building a distributed system which helped in revenue data extraction across various channels and marketplace. This helped in revenue observability and generated valuable insights.- Worked with Selenium, XRPA Client, Open Search , Airflow and Flask. - Revamped the code base to handle parallel processing of jobs and reduce the tech debts. - Worked on architectural improvements and reduced the cost of system by 50$/month. - Led a team of 4 to drive this to completion.

Founding Engineer. All things TechDomain - Fintech: Lending

• Implemented OpenLDAP Services for Vendor Managment• Secured the Interaction between the client and the master node using SSL protocol• API provides mechanism to bind an object to name• Integrated OpenLDAP with The Java Naming and Directory Interface (JNDI),Java API for a directory service that allows Java software clients to discover and look up data and objects.• Provided API for creation of new Hierarchical Directory Structure and Creation of new node.• Implemented a algorithm to search in a directory structure of vendors • Sync with up different vendor for each update and reduced latency in querying data

Development of Delphi API which could be used for creation,manipulation and viewing of XLXS type of Document.

Worked on Openstack and Containers