Vennela Kosanam
Dallas,USA
Summary
Accomplished Python Backend & ML Engineer with expertise in building scalable microservices at BNY. Enhanced fraud detection accuracy to 95% using FastAPI and Flask. Proficient in cloud platforms and containerization, with strong problem-solving skills. Successfully integrated real-time data pipelines, improving system efficiency and reliability.
Overview
5
5
years of professional experience
Work History
Python Backend & ML Engineer
BNY
Connecticut
08.2024 - Current
- Built Python-based backend microservices for fraud detection using FastAPI/Flask, deployed on Vertex AI and Cloud Run.
- Improved fraud detection accuracy from 84% → 95%, reducing false positives by 12% and improving analyst efficiency.
- Implemented FAISS-style vector similarity search to analyze behavioral embeddings for anomaly detection.
- Designed and deployed low-latency serverless inference APIs processing 10M+ monthly transactions.
- Reduced model response time by 30% through pruning, quantization, optimized batching, and caching.
- Integrated Snowflake, PostgreSQL, and BigQuery pipelines with real-time backend services.
- Containerized backend systems using Docker and contributed to CI/CD deployment workflows.
Backend Python Engineer (Volunteer)
Saayam For All
01.2022 - Current
- Engineered geospatial microservices using Python, Flask, AWS Lambda, and DynamoDB.
- Implemented S2 tile indexing with FAISS-style nearest-neighbor logic for real-time volunteer matching.
- Developed REST APIs including /put_user_location, /get_user_location, /find_nearest_volunteers, and availability services.
Python Backend & ML Engineer
Vivma Software Inc.
, India
01.2021 - 12.2022
- Built scalable backend APIs with Python, Flask, and FastAPI to support recommendation and prediction microservices.
- Designed embedding-based similarity matching using PyTorch, boosting user engagement by 23%.
- Developed containerized ML inference pipelines using Docker + Kubernetes with auto-scaling support.
- Built AWS ETL and automated model retraining workflows using EC2, S3, and Lambda.
- Delivered modular, production-ready backend Python services for recommendation, risk scoring, and analytics workloads.
- India
Education
Master of Science - Data Science
University of New Haven
CT12.2024
Bachelor of Technology - Computer Science
Jawaharlal Nehru Technological University
Kakinada05.2021
Skills
- Python and SQL
- FastAPI and Flask
- RESTful APIs
- Microservices architecture
- Database management systems
- FAISS and ChromaDB
- Pinecone embeddings
- Serverless architecture
- Event-driven design patterns
- Cloud platforms (AWS, GCP)
- PostgreSQL and DynamoDB
- MongoDB and Redis
- ETL pipeline development
- Airflow orchestration
- Kafka messaging system
- Containerization with Docker
- Kubernetes orchestration
- CI/CD practices
- Version control with Git
- MLflow tracking
- LangChain integration
- OpenAI API utilization
Projects
Vector Search API (FAISS + FastAPI)
- Built a production-ready vector similarity API using FAISS and FastAPI, supporting embedding generation and sub-50ms retrieval on 100K+ vectors. Deployed using Docker and AWS Lambda.
Distributed Geospatial Matching Service (AWS + S2 Geometry)
- Designed a serverless geospatial engine using AWS Lambda, DynamoDB, and S2 tile mapping to perform real-time proximity matching for thousands of users.
Fraud Detection Inference API (TensorFlow + FastAPI)
- Deployed a TensorFlow fraud detection model as a FastAPI service with batching, caching, and inference-latency optimization. Integrated CI/CD and Docker-based deployments.
Timeline
Python Backend & ML Engineer
BNY
08.2024 - Current
Backend Python Engineer (Volunteer)
Saayam For All
01.2022 - Current
Python Backend & ML Engineer
Vivma Software Inc.
01.2021 - 12.2022
Master of Science - Data Science
University of New Haven
Bachelor of Technology - Computer Science
Jawaharlal Nehru Technological University