Prashant Tyagi

Software Engineer at Carv

Developed a scalable, modular ATS-integration framework using Node.js/TypeScript, RabbitMQ, and Redis that synchronizes job and candidate data bi-directionally across multiple ATS platforms (Greenhouse, Lever, Lift, Workday, and more). Enhanced throughput by 35% and reduced API errors by 40% through schema-driven validation, optimized batching, and robust retry logic—with SLAs enforced by contract-first OpenAPI specifications and automated contract tests. Decreased support load by 30% and cut MTTR by 50% by implementing a unified customer portal that consolidates all integration alerts and tickets into one dashboard. Facilitated the onboarding of 5+ ATS partners in 3 months by translating requirements into technical designs, running end-to-end tests, and coordinating go-lives. Secured 99.9% uptime by configuring Prometheus/Grafana monitoring, setting up Sentry alerts for anomalies, and automating deployments with GitHub Actions, Docker, and Kubernetes—reducing release cycles by 20%

Senior Software Engineer at Virtualworkforce.ai

Led backend architecture design for a greenfield AI-driven workforce platform, establishing microservices boundaries and data contracts for 5K+ daily users. Constructed 20+ RESTful endpoints in Python/Django-DRF with OAuth2 authentication and role-based access control to safeguard sensitive HR data. Containerized services using Docker Compose and streamlined deployments via GitHub Actions and Kubernetes, shortening release-to-production time from 4 hours to 45 minutes. Optimized API performance by 30% through database query enhancements and Redis caching, effectively managing spikes of 1,200 requests per second.

Software Engineer at Metabolic B.V

Engineered core backend services in Go (gin/gRPC) and Python to support over 2M transactions per day for carbon-footprint reporting. Managed GCP deployments using GKE and Cloud SQL (Postgres) to achieve 99.95% uptime and auto-scaling during 5x traffic surges. Optimized data queries by designing a low-latency API layer with eager-load and joined-load techniques, reducing average query time from 450ms to 20ms. Leveraged large language models (LLMs) and the OpenAI API to automate product taxonomy development, boosting classification accuracy by 28%. Integrated event-driven microservices using Pub/Sub and Dataflow, tripling data ingestion throughput. Promoted code quality by enforcing SOLID principles, incorporating MyPy static type checks, and raising unit test coverage from 60% to 92%.

Full Stack Engineer at Hope Technik Pte Ltd

Designed a smart city system utilizing Apache Kafka to process data from IoT devices and sensors, enhancing real-time monitoring of city assets. Implemented cloud-based solutions in an AWS environment with a focus on serverless architecture. Collaborated with product managers and designers to translate user requirements into effective software solutions. Integrated development tools with cross-functional teams to streamline the software development workflow. Developed and maintained CI/CD pipelines to ensure automated testing, frequent releases, and high-quality software.

Software Engineer at Sopra Steria

Developed web APIs in Rust for a microservices architecture, facilitating asset listing, search functionality, and log management, and deployed these APIs using Docker and Kubernetes. Engineered a machine learning model in Python using TensorFlow, deployed via Flask REST API and Django, to recognize objects in images within a Linux environment equipped with an Nvidia K80 GPU. Contributed to frontend development by building responsive SPAs and optimizing component communication through state management using Vue.js, Webpack, and Vuetify.

Software Engineer at Atmus Robotics

Created a facial recognition tool using a wireless camera and a Vue.js based UI, with Node.js middleware for real-time processing. Implemented a system that utilizes 128-bit feature vectors and a Siamese network to perform real-time face matching, supporting applications in security, surveillance, and human behavior analysis.

Software Engineer at DeepPulse Labs

Built an image engine capable of detecting scene types, counting individuals, assessing emotions, quality attributes, pose estimation, and facial landmarks using TensorFlow, OpenCV, and deep learning. Trained a Faster R-CNN model on 12,000 images from a filtered subset of the SUN 2012 dataset to detect 231 distinct objects. Enhanced image context extraction by implementing a sorting algorithm that ranks images based on scored points, thereby presenting the most relevant images. Developed and deployed a machine learning model in TensorFlow for object recognition via a Flask REST API, utilizing Docker and Kubernetes, with supplementary components written in Django.

Software Engineer at Carv

Developed a scalable, modular ATS integration framework using Node.js/TypeScript, RabbitMQ, and Redis to synchronize job and candidate data across multiple ATS platforms (Greenhouse, Lever, Lift, Workday, and more). Enhanced throughput by 35% and reduced API errors by 40% through schema-driven validation, optimized batching, and robust retry logic with contract-first OpenAPI specifications and automated contract tests. Configured monitoring and automated deployments to improve uptime and reduce release cycles.

Senior Software Engineer at Virtualworkforce.ai

Led backend architecture for a greenfield AI-driven workforce platform, establishing microservices boundaries and data contracts for 5K+ daily users. Built 20+ REST endpoints in Python/Django-DRF with OAuth2 authentication and RBAC; containerized services with Docker Compose; streamlined deployments via GitHub Actions and Kubernetes; improved API performance by 30%.

Software Engineer at Metabolic B.V.

Engineered core backend services in Go (gin/gRPC) and Python to support 2M+ daily transactions for carbon-footprint reporting. Managed GCP deployments using GKE and Cloud SQL to achieve 99.95% uptime; optimized data queries with eager-load and joined-load techniques, reducing average query time from 450ms to 20ms. Integrated LLMs/OpenAI API to automate product taxonomy, boosting classification accuracy by 28%.

Full Stack Engineer at Hope Technik Pte Ltd

Designed a smart city system utilizing Apache Kafka to process data from IoT devices and sensors, enhancing real-time monitoring of city assets. Implemented cloud-based solutions in AWS with a serverless focus. Collaborated with product managers and designers to translate requirements into software solutions and maintained CI/CD pipelines for automated testing and frequent releases.

Software Engineer at Sopra Steria

Developed web APIs in Rust for a microservices architecture, facilitating asset listing, search functionality, and log management, and deployed these APIs using Docker and Kubernetes. Engineered a machine learning model in Python (TensorFlow) exposed via Flask/Django to recognize objects in images; contributed to frontend development with Vue.js, Webpack, and Vuetify.

Software Engineer at Atmus Robotics

Created a facial recognition tool using a wireless camera and a Vue.js-based UI, with Node.js middleware for real-time processing. Implemented a system that utilizes 128-bit feature vectors and a Siamese network to perform real-time face matching for security and surveillance applications.

Software Engineer at DeepPulse Labs

Built an image engine capable of detecting scene types, counting individuals, assessing emotions, quality attributes, pose estimation, and facial landmarks using TensorFlow and OpenCV. Trained a Faster R-CNN model on 12,000 images and deployed a Flask REST API for object recognition with Docker and Kubernetes; contributed frontend work with Django.

Software Engineer at Carv

Developed a scalable, modular ATS-integration framework using Node.js/TypeScript, RabbitMQ, and Redis to synchronize job and candidate data across multiple ATS platforms. Improved throughput by 35% and reduced API errors by 40% through schema-driven validation, optimized batching, and robust retry logic with contract-first OpenAPI tests. Built a unified customer portal consolidating alerts and tickets, reducing support load by 30% and MTTR by 50%. Onboarded 5+ ATS partners in 3 months and maintained 99.9% uptime via Prometheus/Grafana, Sentry, and automated deployments with GitHub Actions, Docker, and Kubernetes, cutting release cycles by 20%.

Senior Software Engineer at Virtualworkforce.ai

Led backend architecture for a greenfield AI-driven workforce platform; defined microservices boundaries and data contracts for 5K+ daily users. Implemented 20+ REST endpoints in Python/Django-DRF with OAuth2 authentication and RBAC. Containerized services with Docker Compose; streamlined deployments via GitHub Actions and Kubernetes, reducing release-to-prod time from 4 hours to 45 minutes. Improved API performance by 30% through DB optimizations and Redis caching; managed spikes of 1,200 requests/sec.

Software Engineer at Metabolic B.V.

Engineered core backend services in Go (gin/gRPC) and Python to support over 2M transactions per day for carbon-footprint reporting. Managed GCP deployments using GKE and Cloud SQL, achieving 99.95% uptime and auto-scaling during 5x traffic surges. Optimized data queries with eager-load and joined-load techniques, reducing average query time from 450ms to 20ms. Leveraged LLMs and OpenAI API to automate product taxonomy, tripling data ingestion throughput via event-driven microservices with Pub/Sub and Dataflow. Improved code quality with SOLID principles and MyPy; increased unit test coverage from 60% to 92%.

Full Stack Engineer at Hope Technik Pte Ltd

Designed a smart city system utilizing Apache Kafka to process data from IoT devices for real-time monitoring of city assets. Implemented cloud-based serverless solutions on AWS and collaborated with product managers and designers to translate requirements into effective software solutions. Developed and maintained CI/CD pipelines to ensure automated testing, frequent releases, and high-quality software.

Software Engineer at Sopra Steria

Developed web APIs in Rust for a microservices architecture, enabling asset listing, search functionality, and log management; deployed these APIs using Docker and Kubernetes. Engineered a machine learning model in Python using TensorFlow, deployed via Flask REST API and Django, to recognize objects in images. Contributed to frontend development by building responsive SPAs with Vue.js, Webpack, and Vuetify.

Software Engineer at Atmus Robotics

Created a facial recognition tool using a wireless camera and a Vue.js-based UI, with Node.js middleware for real-time processing. Implemented 128-bit feature vectors and a Siamese network to perform real-time face matching for security and surveillance applications.