Pranjal Kumar
Pranjal Kumar
Work History

Software Engineer 1

Bentley Systems · Pune

Aug 2026 – Present·1 mo

• worked on the backend architecture of iModel Lifecycle Management, focusing on scalable synchronization services, distributed workflows, and Temporal-based orchestration

Associate Software Engineer

Bentley Systems · Remote

Aug 2024 – Jul 2026·1 yr 11 mos

• Owned and built a high-throughput validation microservice in Java/Spring Boot powering compliance checks across 50+ enterprise digital twins; implemented circuit-breaker patterns and exponential backoff, reducing p95 latency from 820ms to 210ms (74% reduction) • Traced intermittent 5xx errors to a race condition in config validation via distributed traces; optimized PATCH endpoint from O(M*N) to O(M+N), cutting response time by 63% and resolving a P1 bug impacting 2,400+ users • Authored design doc and architected a reusable Spring-based network layer, evaluating 3 approaches before selecting circuit-breaker with retry; adopted across 5 microservices, eliminating 90% of duplicate HTTP code • Built a pluggable NLP-to-SQL Node.js package using LangChain and Azure AI, adopted by 3+ teams; automated CI/CD via Azure DevOps reduced release cycle from 2 days to 15 minutes • Engineered a native contour visualization engine, eliminating a 200MB+ VTK dependency; reduced bundle size by 35% and achieved 60fps animations, enabling 1K+ infrastructure engineers to analyze plate stress in-browser • Led architecture and delivery of CDC Client: dynamic rendering,,cmdx parsing, gRPC orchestration; consolidated 3 tools into one app, simplifying workflows for 1K+ structural engineers • Owned operational health of 3 microservices; instrumented with Azure App Insights (p50/p95, error rates), defined SLOs, managed canary rollouts on Kubernetes; reduced MTTD from 45min to under 5min, rollbacks by 60% • Designed a custom RBAC system for Scenes with granular IAM-like policies; collaborated cross-team with Platform and Identity teams to integrate with existing auth infrastructure • Mentored 2 interns through full feature lifecycle from design doc to production deployment; conducted 40+ code reviews focused on scalability patterns and operational readiness

Software Engineer Intern

Bentley Systems · Kolkata, India

Feb 2024 – Aug 2025·1 yr 6 mos

iTwin Services: Validation Team • built POC app using electron for visualizing clashes/overriding on 3d-models/imodels • developed benchmarking API for Clash Detection, reducing analysis time by 85% by creating automated workflow

Evolve Trainee

JMAN Group · Remote

Jan 2024 – Feb 2024·1 mo

• launched a learning platform with <1s TTI and 95+ Lighthouse scores • drafted modular, stateful UI with Tailwind, hooks, and auth guards • created a dynamic admin panel with RBAC, custom Cloudinary media uploads, and dashboard analytics • scraped 5,000+ eyewear SKUs from the company page using browser automation Tech: PostgreSQL, Tailwind, React, Express, Selenium, Azure

Software Lead (Freelance)

R K Trading Co. · Remote

Sep 2023 – Nov 2023·2 mos

• owned the entire lifecycle of RK Trading’s web platform; from design & development to deployment • optimized performance with lazy loading, code splitting, and Lighthouse best practices • handled end-to-end DNS configuration and deployment, ensuring high uptime and fast global resolution Tech: Next.Js, Express, DNS Administration, Cache management, CDN, Technical Leadership, Project Management

Web Development Lead

Google Developer Groups on Campus - CVRGU · Bhubaneswar, Odisha, India

Jul 2022 – Jun 2023·11 mos

• spearheaded 10+ high-impact events: hackathons, tech talks, and hands-on labs with 200+ total attendees • led deep-dive sessions on system design fundamentals, covering scalability, database sharding, caching, and distributed systems • delivered a DevOps essentials track covering containerization, CI/CD pipelines, and deployment strategies Tech: Full-stack development, Technical leadership, DevOps, Software architecture, UI/UX

Full Stack Developer Intern

IEEE Bombay Section · Mumbai

Dec 2022 – Feb 2023·2 mos

• developed 3-way ICO generators, integrating Svelte-Navigator routing • landing page prototype in Figma on a 12-column layout • merging PRs, administering and resolving conflicts in the repo • firebase deployment on top of Svelte framework & led a team of 10 • awarded as top-performing intern with Letter of Recommendation Tech: Svelte, Firebase, Figma, UI/UX, Project management

Flutter Developer

F Square Media · Remote

Sep 2022 – Nov 2022·2 mos

• developed & designed end-to-end cross-platform app for Restaurant aggregation business • utilized the concept of geo-hashing, and developed robust algorithms for backend Tech: Flutter · Figma · GeoHash · UI/UX · Model-View-Controller (MVC) · Project Management · REST APIs

Cyber Security Intern

FireShark · Gurugram, Uttar Pradesh, India

Apr 2022 – May 2022·1 mo

• optimized resource utilization API's by changing the structure of response • worked extensively on reconnaissance techniques • engaged with ciphering(MD5, SHA-256), authentication, salts, cookies • worked hands-on with Linux distributions like Ubuntu and ParrotOS Tech: Vagrant, WSL, VM-Ware, Cybersecurity, Linux, Docker, Bash, CI-CD, Containers, Orchestration

Software Developer (Freelance)

Money Matters · Remote

Jul 2021 – Aug 2021·1 mo

• delivered high-quality, robust production code for a Chartered Accountant Firm • provided leadership as a distinguished developer covering multiple domains in the project Tech: React, Express, SEO, Google Analytics, Pipelines, CI/CD

Jobs Content
Production Incident: When good rate limits, block good users | Pranjal Kumar (@PranjalKumar__) on X

Production Incident: When good rate limits, block good users | Pranjal Kumar (@PranjalKumar__) on X

A production incident taught me that a good-looking rate limit can still be wrong. We had a rate limiter that looked fine in code and terrible in prod...

problem solving
system design
architecture
Production Incident: When good rate limits, block good users | Pranjal Kumar (@PranjalKumar__) on X
Story of AWS S3: "You don’t buy the fastest disk. You buy so many slow ones that they behave like a supercomputer"💪🏻 | Pranjal Kumar (@PranjalKumar__) on X

Story of AWS S3: "You don’t buy the fastest disk. You buy so many slow ones that they behave like a supercomputer"💪🏻 | Pranjal Kumar (@PranjalKumar__) on X

Story of AWS S3: "You don’t buy the fastest disk. You buy so many slow ones that they behave like a supercomputer💪🏻" I thought S3 must be SSD-backe...

aws
problem solving
scaling
Story of AWS S3: "You don’t buy the fastest disk. You buy so many slow ones that they behave like a supercomputer"💪🏻 | Pranjal Kumar (@PranjalKumar__) on X
Configuring robust secure databases | Pranjal Kumar (@PranjalKumar__) on X / Twitter

Configuring robust secure databases | Pranjal Kumar (@PranjalKumar__) on X / Twitter

Your “private data” is only as private as the weakest database that stores it. In distributed systems, we trade off consistency. In digital life, we t...

database
security
Configuring robust secure databases | Pranjal Kumar (@PranjalKumar__) on X / Twitter
Optimizing cloud infrastructure of UK's tax authority (HMRC) | Pranjal Kumar (@PranjalKumar__) on X/Twitter

Optimizing cloud infrastructure of UK's tax authority (HMRC) | Pranjal Kumar (@PranjalKumar__) on X/Twitter

Running a national-scale service from us-east-1 isn’t just a compliance risk, but also a latency nightmare. Every tax query literally crosses the Atl...

cloud computing
aws
system design
scaling
Optimizing cloud infrastructure of UK's tax authority (HMRC) | Pranjal Kumar (@PranjalKumar__) on X/Twitter
Space-filling curves like Hilbert map multi-dimensional data into 1D | Pranjal Kumar (@PranjalKumar__) on X/Twitter

Space-filling curves like Hilbert map multi-dimensional data into 1D | Pranjal Kumar (@PranjalKumar__) on X/Twitter

Space-filling curves like Hilbert map multi-dimensional data into 1D while preserving locality. That’s why databases and caches love them Hilbert curv...

space filling curve
algorithms
fractals
Space-filling curves like Hilbert map multi-dimensional data into 1D | Pranjal Kumar (@PranjalKumar__) on X/Twitter
Don't make users wait for a task, instead use Job ID and Status | Pranjal Kumar (@PranjalKumar__) on X/Twitter

Don't make users wait for a task, instead use Job ID and Status | Pranjal Kumar (@PranjalKumar__) on X/Twitter

Friday wisdom: Don't wait for long-running jobs. Instead, return quickly with a Job ID, that can be used to track the progress of the job.

learnings
user experience
Don't make users wait for a task, instead use Job ID and Status | Pranjal Kumar (@PranjalKumar__) on X/Twitter
Get the boundaries of your VPC right | Pranjal Kumar (@PranjalKumar__) on X/Twitter

Get the boundaries of your VPC right | Pranjal Kumar (@PranjalKumar__) on X/Twitter

One of my biggest lessons: VPC design is like API design. Get the boundaries wrong (public vs private subnets) and you’re fighting leaks forever

networking
subnets
infrastructure
Get the boundaries of your VPC right | Pranjal Kumar (@PranjalKumar__) on X/Twitter
Reads to Writes ratio is heavily skewed in Softwares | Pranjal Kumar (@PranjalKumar__) on X/Twitter

Reads to Writes ratio is heavily skewed in Softwares | Pranjal Kumar (@PranjalKumar__) on X/Twitter

Working with Legacy Senior Engineers makes you realise that "if you want your code to be easy to write, make it easy to read". The Reads to Writes rat...

learnings
system design
Reads to Writes ratio is heavily skewed in Softwares | Pranjal Kumar (@PranjalKumar__) on X/Twitter
Google's history of building revolutionary because of their scale | Pranjal Kumar (@PranjalKumar__) on X/Twitter

Google's history of building revolutionary because of their scale | Pranjal Kumar (@PranjalKumar__) on X/Twitter

Every Google tool exists because they hit scaling walls 5 years before anyone else. That’s both their advantage and their curse. Borg before Kubernete...

software engineering
architecture
system design
Google's history of building revolutionary because of their scale | Pranjal Kumar (@PranjalKumar__) on X/Twitter
Google Search API new limits restrict to 10 search results | Pranjal Kumar (@PranjalKumar__) on X/Twitter

Google Search API new limits restrict to 10 search results | Pranjal Kumar (@PranjalKumar__) on X/Twitter

Most AI startups don’t have their own web crawlers; they rely on Google’s index as a proxy for the open web. That dependency just broke with Google's ...

system design
scaling
architecture
Google Search API new limits restrict to 10 search results | Pranjal Kumar (@PranjalKumar__) on X/Twitter
Communities

GDG New Delhi

59274 members

GDG Cloud New Delhi

52754 members

AI Community Delhi

5655 members

CDN Commudle Developer Network

6843 members

TensorFlow User Group Ghaziabad

9197 members

Google Developers Group Siliguri

2176 members