Sagar Pasrija

Software Engineer

South Delhi, Delhi, India7 yrs experience

Highly StableAI Enabled

Key Highlights

Led migration to Golang, reducing latency by 70%.
Developed high-performance recommendation platform.
Created config-driven Inference Engine for ML models.

Stackforce AI infers this person is a Backend-heavy SaaS Engineer with expertise in high-traffic systems.

Contact

Skills

Core Skills

AwsGolangMachine LearningGenerative Ai DevelopmentJava DevelopmentDevopsBigtablePython

Other Skills

Amazon Web Services (AWS)Data AnalysisSoftware Developmentn8nMavenMySQLMicroservicesKubernetesNoSQLSystem MonitoringBigQueryComputer ScienceCommunicationRecommender SystemsAWS Lambda

About

SDE2 with the experience of leading two teams of engineers to create stable and scalable products in diverse domains. 6+ years of hands-on technical experience in building critical engineering systems such as large scale consumer products at Nykaa and Magicpin.Diverse experience in different systems (user facing high traffic systems, search and transactional systems).Strong foundation in Data Structures and Algorithms. Comfortable with coding in Golang, Python, Java-Spring. Industry level working experience with Aerospike, AWS, Bigtable, Bigquery, Cassandra, DAX, DocumentDB, DynamoDB, GCE, GKE (Kubernetes), Lambda, Kafka, MongoDB, nginx, Redis, Solr, SQL, SQS. Familiar with ML, Neural Networks, NLP.Journal publications in leading conferences.Bachelor of Technology (Netaji Subhas Institute of Technology) with research internship at Freie Universität, Berlin, Germany.

Experience

7 yrs

Total Experience

5 yrs 4 mos

Average Tenure

1 yr 9 mos

Current Experience

Nykaa

SDE2

Sep 2024 – Present · 1 yr 9 mos · Gurugram, Haryana, India

Led the migration of a high-traffic, user-facing Python widget service (≥200k RPM) to Golang, achieving 70% lower latency, 30% infra cost reduction, improved autoscaling, and 5-7 min faster builds through better resource utilisation.
Worked on high-performance recommendation platform delivering user, cohort, and item-based recommendations within 30ms.
Designed and developed a config-driven Inference Engine from scratch: a self-serve platform enabling data science teams to code, test, deploy, version, and monitor ML models with full experimentation support.
Implemented dynamic widget ranking, enabling real-time filtering and reordering based on user-level metrics to boost CTR and conversion.
Worked with n8n, LLMs and GenAI to build a generic support bot using a RAG architecture, helping ops and product teams debug flows, resolve setup issues, and answer natural-language queries.
Took ownership of multiple microservices and contributed to key architectural decisions across the platform.

Generative AI DevelopmentAmazon Web Services (AWS)AWSGolang

Magicpin (samast technologies)

3 roles

Technical Lead

Jun 2022 – Oct 2024 · 2 yrs 4 mos

Managed multiple workstreams across two teams of 6-8 developers each and learned how to be a project manager along with people manager. Led the business use case requirements and built, tested, reviewed and deployed production ready APIs in Golang, Java with Spring and Python.
Discovery: Developed and managed the microservice from scratch in Golang. Along with a Content-Management System (CMS), discovery enabled the content team to make pages and widgets live directly on the app and web.
Wallet: Working on and managing Magicpin's transactional system.
Managed Magicpin's infrastructure including but not limited to load balancers, instance groups, K8 workloads, solr, aerospike, redis, cassandra and bigtable.
Migrated Golang, Java and Python based microservices, kafka consumers and stateful applications (such as solr and aerospike) from GCE to GKE.
Setup CI/CD pipeline for multiple projects through bitbucket, thereby automating deployment of different workloads directly to K8.

MavenMySQLGolangJava Development

Senior Software Engineer

Promoted

Jun 2021 – Jun 2022 · 1 yr

Segmentation Service: Built using Golang, Bigtable, BigQuery and K8 cronjobs, it enables teams to create dynamic user segments and use them across different services to control notifications, widget/offer visibility, coupon code applicability and much more.
Optimised code flows and brought down latencies of critical APIs from ~3sec to 200ms-700ms.

NoSQLSystem MonitoringGolangBigtable

Software Engineer

Jun 2019 – Jun 2021 · 2 yrs

De-Duplication Service: Developed de-duplication service using minhash and LSH in Python successfully de-duplicating more than 3M merchants with over 92% accuracy.
Merchant Recommendation: Applied Unsupervised Learning and BigQuery to put in place a self training merchant recommendation system.

NoSQLComputer SciencePython