Work Experience

Oct 2025 – Present · Pune, India AI-driven fintech company building intelligent financial automation products

Developed a centralized budgeting and planning module using FastAPI and PostgreSQL, generating P&L, Trial Balance, Chart of Accounts, and procurement planning outputs for streamlined financial reporting.
Built a multi-source ERP data ingestion layer integrating QuickBooks, Zoho (OAuth), and Tally; orchestrated ETL workflows via Airflow with Cloud Run-based extraction and dbt transformations — reducing manual data preparation effort by 40%.
Engineered Financial Insights to generate AI-driven insights from standardized ERP data via the AskChamp chatbot pipeline; built scheduling and user/group-based notification workflows for proactive financial monitoring.
Enhanced AskChamp by integrating an AI SQL agent with an MCP-based orchestration layer for LLM-driven query orchestration, hybrid retrieval, and pinned-query workflows across 10+ knowledge bases; deployed on Vertex AI (GCP), improving query accuracy by 30%.
Built LLM evaluation and reliability systems including LLM-as-a-Judge regression test suites, feedback-driven knowledge base update pipelines, and RBAC-enabled admin platforms for user access management.

May 2024 – Jun 2025 · Bloomington, IN, USA

Engineered 5+ ETL pipelines using Azure Data Factory to ingest large-scale public education and healthcare data (CSV, XML, JSON) from 70+ counties into Blob Storage, reducing data extraction errors by 30%.
Built predictive models (Random Forest, XGBoost, ARIMA) achieving 85% accuracy on post-graduation salary and patient volume forecasts; delivered Tableau dashboards tracking 10+ KPIs with geographic and demographic filters to drive policy decisions.

May 2023 – Aug 2023 · Troy, MI, USA

Deployed 5+ ETL pipelines to process large-scale time-series sensor data from 100+ injection molding machines, ingesting 80+ GB into InfluxDB for real-time monitoring of temperature, pressure, and flow metrics.
Built classification models (Logistic Regression, Decision Trees, Random Forest) to predict chiller tank overflow risks, reducing malfunctions by 10%; built Grafana dashboards with automated alerts and CI/CD deployment.

Aug 2021 – May 2022 · Bengaluru, India

Constructed scalable AWS pipelines (Glue, PySpark, S3) to ingest and process 3D radar point-cloud and JSON data from 10+ REST APIs, improving data integration efficiency by 25%.
Developed an object classification model using PointNet++ in Databricks for vehicle and pedestrian detection, improving ADAS speed control responsiveness by 20% and model stability by 10%.

Jan 2021 – May 2021 · Pune, India

Developed an ASR system for 100+ audio recordings in AWS S3 using Python and Librosa for MFCC feature extraction.
Improved ASR accuracy to 90% via transfer learning on a DeepSpeech model with custom lexicon tuning and hyperparameter optimization.