Chetan Pandey
Sr. Software Engineer
Building scalable distributed systems and robust backend infrastructure. Specializing in microservices, AI & LLM integration, GenAI workflows, and cloud-native solutions.
About Me
Building systems that scale & endure
Software Engineer with 6+ years of experience building high-performance distributed systems across cloud, Web3, and enterprise platforms. I work extensively with Golang, Kubernetes, NodeJS, Python, and Microservices, with a strong focus on AI integration, LLM, GenAI, and Baremetal Infra.
I've architected fault-tolerant systems, implemented chaos engineering practices, integrated real-time AI speech and LLM services, and optimized distributed databases for reliability and performance across AWS Cloud and Linux environments.
Beyond code, I contribute to open-source projects, mentor junior engineers, and stay current with the latest in distributed computing, AI/ML tooling, and blockchain infrastructure.
Productivity → Performance → Reliability → Scale
Technical Skills
Tech Stack & Expertise
Languages
Go-first stack for high-performance services and tooling
AI & GenAI
LLM integration, real-time speech AI, and GenAI workflows
Backend & APIs
REST, event-driven, and real-time API patterns at scale
Infrastructure
Container orchestration, IaC, and zero-downtime deployments
AWS Cloud
Multi-region serverless and containerised architectures
Databases
Relational, NoSQL and in-memory stores tuned for scale
Messaging
High-throughput event streaming and async message queues
Web3 & Blockchain
Decentralised infra, DeFi protocols and smart contracts
Frontend
Modern web interfaces with React ecosystem and design systems
System Design
Fault-tolerant distributed systems with chaos engineering
Work History
Professional Experience
- Designed and developed Voice Activity Detection (VAD) for AI-Agents bot, improving speech detection accuracy by ~30% with ~1ms latency
- Integrated AI speech and LLM services — ElevenLabs, Deepgram, AssemblyAI, and OpenAI (ChatGPT) for real-time synthesis, transcription, and conversational AI workflows
- Designed and integrated noise suppression libraries improving speech detection in noisy environments with ~12–15ms latency
- Optimized Java, Go, and Python applications using AWS Aperf, improving throughput by ~30% without major architectural changes
- Performed load testing of AI-Agents using Cekura, identifying bottlenecks and validating scalability under high traffic
- Conducted performance benchmarking and tuning across microservices, reducing latency and improving system scalability
- Building Decentralized (Web3) infrastructure to meet increasing GPU demand in the cloud ecosystem
- Designed unique network tunneling application in Golang — backbone of Spheron Fizz
- Developed provider Kubernetes operator handling blockchain interaction and deployment lifecycle end-to-end
- Implemented AI Jobs and Inferencing on Fizz Node achieving $1M ARR
- Integrated CopperX, scaling to $50K monthly revenue on SaaS
- Cluster monitoring using Prometheus, Grafana, and OpenTelemetry
- Built CI/CD pipeline for one-click installation across multiple projects
- Redesigned monolith voice backend to microservices improving availability and scalability
- Developed Kubernetes controller and operator in Golang for auto-provisioning services
- Designed Real-Time Call Transcription with ObserveAI integration
- Optimized Golang microservices using pprof, enhancing concurrency and minimizing GC
- Designed custom logic-based router in TypeScript for complex call-routing scenarios
- Developed microservices in Golang using gRPC and HTTP with Kafka/RabbitMQ message buses
- Designed Uploader 2.0 from ideation to release, resulting in 50× faster uploads
- Migrated voice microservices from EC2 to Kubernetes (EKS)
- Built XML-based call flow interpreter and event handler for telephony systems
- Developed REST APIs for telephony system backend in Golang using Echo Framework
- Developed automation scripts using Python and Selenium for MCA data scraping
- Optimized Elasticsearch to query companies index with 1.5M+ documents
- Worked on TensorFlow-based captcha solver for automation scripts
- Parsed and stored scraped data in MySQL database with optimized queries