
Priyanshu Sekhar
Verified Expert in Engineering
Software Developer
Bengaluru, Karnataka, India
Toptal member since May 5, 2021
Priyanshu is a hands-on engineering architect with 10+ years across the full SDLC. His expertise spans advanced Java, real-time distributed systems—with energy trades streamed at around 1 million/sec on Apache Flink/MSK—and production AI. He has led engineering teams, shipped a WhatsApp AI agent serving 10,000 users, and built an in-house platform where LLMs auto-diagnose production errors and coding agents open the fix. He architects ambitious systems and stays deep in the code that ships them.
Portfolio
Experience
- React - 7 years
- Java - 7 years
- Amazon Web Services (AWS) - 6 years
- Next.js - 4 years
- Python - 4 years
- AI Agent Orchestration - 3 years
- Large Language Models (LLMs) - 3 years
- AI Chatbots - 2 years
Preferred Environment
Java, Python, Amazon Web Services (AWS), React, Next.js, Large Language Models (LLMs), AWS Bedrock AgentCore, AI Chatbots, AI Agent Orchestration, AI Agents
The most amazing...
...product I've built takes 10,000+ pharmacy customers from medicine search to a paid order entirely in chat, across 50+ stores.
Work Experience
Senior Software Consultant
TP ICAP
- Architected and built an enterprise-grade streaming platform delivering real-time oil/energy trade data to institutional clients via WebSocket, processing roughly a million trades/second with sub-sec latency using Apache Flink, MSK, and API Gateway.
- Designed an Apache Flink application on AWS Kinesis Data Analytics consuming cross-account MSK topics from Data Warehouse, implementing 5-minute tumbling window aggregations with exactly-once semantics and checkpoint-based recovery.
- Built Spring Boot microservices on EKS, managing WebSocket connections with Okta JWT authentication, per-client Kafka topic routing, and dynamic filtering, supporting 100+ concurrent external clients with client-specific entitlements.
- Provisioned complete AWS infrastructure using Terraform, including MSK clusters with IAM authentication, EKS deployments with Helm, API Gateway WebSocket APIs, Lambda authorizers, cross-account IAM roles, and DynamoDB.
- Built scheduled ETL pipeline processing 100K+ daily market records cached in Redis (fed by MSK) into client-specific CSV snapshots, delivering to S3 and SMB shares with CRC32 validation, Snowflake audit logging, and fault-tolerant recovery mechanism.
Head of Engineering
LokamHealth
- Architected a healthcare platform (FastAPI, Next.js, React Native, Flutter) serving 10,000+ users across 50+ pharmacies with order tracking, prescription validation, inventory with expiry management, and Zoho Books integration.
- Built an AI observability platform, LLM auto-triage that fingerprints errors, RCA and fixes, a pgvector "seen-before" memory, deploy correlation, and an MCP server that lets AI coding agents pull real production issues and open fixes under human-gated review.
- Designed task management with GPS verification and photo uploads for rack optimization, leave tracking, and employee monitoring via Lambda/EventBridge—87% cost reduction vs. polling.
- Created PWA and React Native apps for self-ordering with AI-powered generic medicine recommendations using OpenAI embeddings, cart management, and real-time tracking via SSE.
- Built and shipped a WhatsApp conversational-commerce agent (LangGraph orchestration on Amazon Bedrock) that takes customers from medicine search to placed order entirely in chat—live across 50+ pharmacies for 10,000+ users, handling 5,000 messages per day.
- Designed a human-in-the-loop supervision layer over the agent, real-time sentiment scoring, 1-click operator takeover, a pending-response review queue, and layered memory—so staff supervise and correct AI replies without derailing the conversation.
- Migrated the entire LLM stack to Amazon Bedrock behind a universal fallback provider with graceful degradation, keeping every downstream classifier available through provider outages; re-embedded the catalog on Titan for kNN generics recommendations.
- Stood up production-grade agent observability: self-hosted Langfuse tracing (PII-masked) plus a PostgreSQL and pgvector read-model fed by a Redis outbox to an in-VPC worker, making agent behavior measurable and debuggable on real conversations.
Head of Engineering
Revolutio
- Designed cloud-based architecture for SaaS platform from scratch. Optimized costs for low-volume clients using serverless, event-driven architecture. Achieved infrastructure stack isolation for high-volume clients using multi-tenancy design patterns.
- Developed generative AI assistants for information retrieval and task automation. Leveraged step functions to build agents integrating multiple assistants, enabling a multistep approach to handling complex queries and tasks.
- Increased customer productivity by building plugins for Outlook, Word, and Slack to import relevant data. Reduced onboarding time by 30%. Set up CI/CD pipelines achieving <1% failure rate.
Senior React Native Engineer
Gartner - Peer Insights
- Participated in the release of the Gartner Peer Community, a platform for peer-driven business decisions. Added web pages, rebranded existing pages, and synchronized updates across iOS and Android.
- Took responsibility for the web release of channels in the Gartner Peer Community, allowing users to post and watch content in specific channels.
- Served as one of three leads responsible for the release of new products split from the Gartner peer platform to peer insights and peer community. This involved several updates in iOS, Android, web, and mobile web.
Coding Tech Lead
Buzz
- Built a desktop web app for remote collaboration using Flutter and AWS, including floating widgets for notifications and video, real-time syncing of user calendars and slack statuses, and teams to organize the workplace.
- Designed and built a new feature for creating teams, presenting organization hierarchy, and allowing people to join meetings and show the statuses of other team members. Most clients heavily demanded this feature; it was delivered in two months.
- Acted as a team lead, established processes for backlog grooming, scheduling releases, and updates, streamlining communication around reporting and fixing bugs, and creating documentation for all features and roadmap.
AWS Expert
Com2 Media LLC
- Developed a journaling app using React Native, allowing users to write notes, follow users, and get notified when a new journal is entered by people they follow.
- Built the client's app from scratch, integrating with MongoDB for document storage and Lambda for triggering notifications.
- Established the back end infrastructure as a code and connected MongoDB events with AWS services using EventBridge to get real-time updates. Added analytics using Pinpoint and SNS.
Lead Developer
Mymory
- Designed multimodal data ingestion pipelines to capture information in a unified format from multiple 3rd-party apps and real-time sources. Built Android, iOS, and web apps for customers to store, query, and organize information.
- Designed and developed the complete back-end infrastructure built on a serverless framework using AWS Lambda, authenticating via Cognito, and using DynamoDB and S3 to store data and objects.
- Integrated state-of-the-art machine learning tools for object detection in videos, OCR in-text images, data extraction from PDFs, and finding relevant results from searches based on natural language.
Senior Back-end Engineer
Digibuild
- Designed and built several workflows for order fulfillment, including invoices, quote, purchase orders, and more.
- Built the testing framework for the back end and front end to verify database migrations, API-level unit tests, and screen UI testing, integrating those tests into the deployment pipeline.
- Modularized the front-end code to build a more scalable architecture. Optimized GraphQL and REST back-end API queries, improving latency and cost.
- Integrated with several AWS services for messaging, notifications, serverless back end, and scheduled jobs, improving the usability of Digibuild software.
- Designed the blockchain framework to store transactions on the Ledger using AWS-managed blockchain.
Senior Software Development Engineer
Amazon.com
- Designed and built a batch ingestion pipeline to process over a billion knowledge facts, generate additional lexical properties with a 50% reduction in execution times, and 25% less storage space.
- Optimized big data workflows and ETL jobs to achieve over twofold increase in efficiency and 85% cost reduction, while simultaneously scaling data ingestion from 100 million to over two billion facts per day.
- Developed natural language models, expanded the knowledge graph database of Alexa, to understand and respond to a broader spectrum of queries with minimal latency.
Software Engineer 2
Intuit
- Built a company migrator tool to automate the cloning of all client-related information from one QA environment to another.
- Designed the API migration of services from in-house providers to AWS.
- Audited and fixed security bugs in microservices while transitioning to AWS.
Co-founder, Tech Architect
Naradd
- Designed the architecture and back-end implementation of the Naradd client portal. Built features for client onboarding, managing company information, and team-member access. Integrated APIs to send SMS, WhatsApp, email, and push notifications.
- Built query and analytics over large data files (uploaded by clients) using AWS Athena and Glue.
- Designed and implemented a pipeline to profile users and ensured optimal engagement across several communication mediums.
- Contributed to the initial development of the Naradd Android app. Integrated features for authentication, category selection, and subscribe and unsubscribe to companies.
- Built the Naradd web app for iOS users with integration to WhatsApp for transactional communications.
- Managed software development interns and assisted them in onboarding to web and app projects.
Software Engineer
PayPal
- Designed and built a microservice for reading transactions instantly from several domains, aggregating and storing to a unified database to be later shown up in the user's activity page. Reduced the ETA of read-to-write by 5x, down to 2 seconds.
- Developed a data reconciliation pipeline using Kafka and PayPal internal messaging queues. Dramatically reduced the messages in the exception queue and improved efficiency in data correction and storage to five times reliability.
- Built a patching tool by attaching the pipeline with a microservice for on-demand data correction. Integrated that with PayPal NFS for easy and fast upload of patch files, facilitating teams in maintaining data consistencies.
Open Source Contributer, GSoC
Apache OFBiz
- Integrated GenApp software with Apache Airavata for executing computational jobs on distributed resources.
- Built a Qt interface for the GenApp software to enable computation from mobile platforms.
- Extended Apache Airavata to include additional monitoring and workflow APIs.
Experience
SAGE — The Creator Operations Brain
https://sage.zephyrcode.liveFORGE — Precision Body Transformation OS
https://forge-app.zephyrcode.live/welcomeThe Arena — Competitive Engineering Puzzles for the AI Era
https://arena.zephyrcode.liveMARGIN — Notes in the Margin of the Web
https://margin.zephyrcode.liveSAMHITA — Prose as Code
https://samhita.zephyrcode.liveEducation
Bachelor's Degree in Electrical Engineering
Indian Institute of Technology, Kharagpur - West Bengal, India
Skills
Libraries/APIs
React, REST APIs, Slack API, Node.js, Amazon EC2 API, WhatsApp API, AWS Amplify
Tools
Amazon CloudFront, Expo, Terraform, AWS Glue, Amazon Elastic Container Service (ECS), Amazon EKS, Amazon Transcribe, Amazon Cognito, AWS Step Functions, AWS IAM, AWS CloudFormation, Flink, Amazon ElastiCache, AWS Key Management Service (KMS)
Languages
Java, Python, TypeScript, JavaScript, Solidity, GraphQL, Dart, CSS, Go, Java 8
Frameworks
Flutter, Spring, Spring Boot, React Native, Django, Spark, Next.js, Redux, Tailwind CSS, Material UI, Angular, Qt 5, Airavata
Paradigms
REST, Agile, Test-driven Development (TDD), Event-driven Architecture
Platforms
Firebase, Amazon Web Services (AWS), AWS Lambda, Google Cloud Platform (GCP), Amazon EC2, Apache Kafka, Blockchain, Ethereum, Music Streaming, FlutterFlow, iOS
Storage
Amazon S3 (AWS S3), MySQL, Cloud Firestore, NoSQL, JSON, PostgreSQL, MongoDB, Realm, Amazon DynamoDB, Elasticsearch, Redis
Other
Software Development, APIs, Back-end, Cloud, Large-scale Distributed Systems, Distributed Systems, Technical Leadership, Software Architecture, Firebase Cloud Functions, API Integration, Full-stack, OpenAI, Architecture, FastAPI, AI Integration, Architectural Design, Fractional CTO, Scalable Platforms, AI-enabled Applications, Communication, Orchestration, Agentic AI, Front-end, CI/CD Pipelines, RESTFul APIs, Infrastructure as Code (IaC), Amazon RDS, Data Architecture, Artificial Intelligence (AI), Generative Artificial Intelligence (GenAI), API Gateways, Enterprise Architecture, AI Agent Orchestration, AI-generated Code, API Design, Large Language Models (LLMs), Software Engineering, AI Chatbots, Software Design, SaaS, SaaS Design, Audio, Videos, Streaming, Optical Character Recognition (OCR), WhatsApp Business, Knowledge Graphs, Advisory, Supabase, AI Agents, Terraform Cloud, Amazon Comprehend, Pinpoint, Blockchain as a Service, AWS Bedrock AgentCore, AWS App Runner, DocumentDB, Data Visualization
How to Work with Toptal
Toptal matches you directly with global industry experts from our network in hours—not weeks or months.
Share your needs
Choose your talent
Start your risk-free talent trial
Top talent is in high demand.
Start hiring