Souvik Roy

CEO

Bangalore Urban, Karnataka, India7 yrs 2 mos experience

AI ML PractitionerHighly Stable

Key Highlights

Expert in Data Engineering and AI solutions.
Led orchestration migration improving workflow efficiency.
Architected data automation saving thousands of man-hours.

Stackforce AI infers this person is a Fintech and Environmental Tech specialist with strong AI and Data Engineering expertise.

Contact

Skills

Core Skills

Data EngineeringAi DevelopmentProcess AutomationData ScienceComputer Vision

Other Skills

PythonAzure DevOpsApache AirflowAsset ControlShell ScriptingPower BIAlteryxSQLFlaskMachine LearningKerasTensorFlowAI AgentsStrategyTechnology Consulting

About

I build production-ready solutions for a financial data management system. Currently, I am working at Standard Chartered, developing solutions on the bank's central system for distributing instrument reference data (Asset Control) to critical downstream systems. Focus: Data Engineering and Agentic AI. Tech Stack: Python, FastAPI, Celery, RabbitMQ Domain: Finance, Asset Control (FE, Consolidation, Normalization) SQL, Azure DevOps, Airflow, Databricks Follow me for insights on AI, software architecture, and life in general.

Experience

7 yrs 2 mos

Total Experience

2 yrs 9 mos

Average Tenure

1 yr 6 mos

Current Experience

Standard chartered

Lead Developer

Nov 2024 – Present · 1 yr 6 mos · Bangalore Urban, Karnataka, India · Hybrid

Managed the end-to-end lifecycle of Instrument Reference Data, ensuring accurate distribution of Price and Static data to downstream Trading and Regulatory reporting teams.
Developing an agentic AI system for solution development with MCP servers for confluence, ADO and other sub-agents using claude sonnet in GHCP agent mode.
Designed data onboarding workflows for complex asset classes, integrating direct feeds from Bloomberg and other vendors into the bank's centralized Asset Control repository.
Leading the "Control-M to Airflow" Orchestration Migration, redesigning complex dependency chains for Reference Data workflows to improve observability and fault tolerance.
Engineered CI/CD pipelines using Azure DevOps, reducing release cycle times and enforcing strict code governance for financial data models.
Tech Stack: Asset Control (Prime EDM), Apache Airflow, Python, Linux, Oracle, Azure DevOps, Shell Scripting.

PythonAzure DevOpsApache AirflowAsset ControlShell ScriptingData Engineering+1

Eclerx

Process Manager | Data Engineering

Sep 2020 – Nov 2024 · 4 yrs 2 mos · Remote

Architected an end-to-end data automation ecosystem for a Tier-1 Investment Bank, replacing legacy manual workflows with Python, Power BI and Alteryx workflows. This architectural overhaul resulted in 2,000+ annualized man-hours saved.
Optimized SQL queries and data models, reducing processing overhead by 30% and ensuring SLA compliance for critical bank reporting.
Developed a localized "Control Tower" architecture using Power BI to monitor ETL job status and data lineage, resulting in a 25% reduction in downtime and data anomalies.

PythonPower BIAlteryxSQLData EngineeringProcess Automation

Indian institute of technology gandhinagar

Research Assistant | Computational Sustainability & AI

Jan 2019 – Jul 2020 · 1 yr 6 mos · Gujarat, India

Architected a Real-Time IoT Monitoring System: Developed a full-stack geospatial dashboard using Python (Flask) and Leaflet.js to ingest and visualize live telemetry from distributed PM2.5 sensors, enabling high-resolution pollution hotspot analysis.
Engineered Spatiotemporal ML Pipelines: Curated a pan-India air quality dataset by performing complex spatiotemporal alignment (collocation) of Satellite Aerosol Optical Depth (AOD) and meteorological data to solve ground-sensor sparsity.
Developed Ensemble Prediction Models: Benchmarked Random Forest and Ridge Regression algorithms with Time-Series Cross-Validation to estimate surface-level PM2.5, successfully modeling seasonal variance across diverse Indian geographies.
Sensor Calibration & Statistical Analysis: Conducted field deployment and regression-based calibration of low-cost IoT sensors against reference-grade equipment to validate data integrity for construction site exposure studies.

PythonFlaskMachine LearningAI DevelopmentData Science

Variable energy cyclotron centre (vecc)

Computer Vision Research Intern

May 2017 – Jul 2017 · 2 mos · Kolkata, West Bengal, India · On-site

Engineered a novel feature extraction pipeline that transformed raw network packet captures (PCAP) into 2D image representations, enabling the application of Computer Vision algorithms to cybersecurity anomalies.
Benchmarked State-of-the-Art (SOTA) CNN architectures (ResNet, VGG16, GoogLeNet) to classify malicious botnet traffic vs. benign activity, analyzing the trade-offs between model depth and classification accuracy.
Curated a synthetic cybersecurity dataset simulating diverse botnet attack vectors, validating model robustness against zero-day network signatures.
Tech: Python, Keras/TensorFlow, OpenCV, TCP/IP Protocols, Convolutional Neural Networks.

PythonKerasTensorFlowComputer Vision