███████╗██╗ ██╗ █████╗ ███████╗██╗ ██╗ █████╗ ███╗ ██╗██╗ ██╗
██╔════╝██║ ██║██╔══██╗██╔════╝██║ ██║██╔══██╗████╗ ██║██║ ██╔╝
███████╗███████║███████║███████╗███████║███████║██╔██╗ ██║█████╔╝
╚════██║██╔══██║██╔══██║╚════██║██╔══██║██╔══██║██║╚██╗██║██╔═██╗
███████║██║ ██║██║ ██║███████║██║ ██║██║ ██║██║ ╚████║██║ ██╗
╚══════╝╚═╝ ╚═╝╚═╝ ╚═╝╚══════╝╚═╝ ╚═╝╚═╝ ╚═╝╚═╝ ╚═══╝╚═╝ ╚═╝
⚙️ auto-generated nightly by GitHub Actions — see setup below
apiVersion: v1
kind: Engineer
metadata:
name: shashank-chakraborty
location: bengaluru-in
status: fourth-year-student
spec:
focus: [backend-engineering, devops-automation, cloud-security]
currentlyBuilding: infrastructure-drift-detection-engine
personality: >
Will rebuild a working CI/CD pipeline for a 6-minute improvement.
Reads incident postmortems for fun. Cannot leave a sequential
pipeline alone — will parallelize it out of principle.
status:
uptime: "since 2023"
lastIncident: "shipped instead of sleeping"
reliability: "measured, not assumed"graph TD
A[Shashank] --> B[Backend APIs]
A --> C[DevOps & Infra]
A --> D[Cloud Security]
B --> B1[FastAPI / Flask / Node.js]
B --> B2[Async pipelines - Celery / Redis]
C --> C1[Docker + Kubernetes + HPA]
C --> C2[CI/CD - GitHub Actions]
C --> C3[Observability - Prometheus / Grafana]
D --> D1[AWS IAM graph analysis]
D --> D2[Privilege escalation detection]
* a1b2c3d (SystemCraft) parallelize CI/CD pipeline
| → 10min sequential → under 4min, fixed artifact-integrity bug too
* 9f8e7d6 (SystemCraft) k6 load test @ 500 concurrent users
| → p95 latency 3.33s → 861ms (-74%), throughput +161%
* 4c5d6e7 (Shadow Permission Analyzer) BFS/DFS on AWS IAM graph
| → detects multi-hop privilege escalation chains in Neo4j
* 2b3c4d5 (DocuFlow) event-driven invoice pipeline
| → async workers cut processing time by 40%
* 1a2b3c4 (Hackman) 24hr national hackathon, team ClarityAI
→ top 10, agentic decision-intelligence tool
SystemCraft — real-time system design simulator (live · repo)
Built to answer one question: does my infra work under load, or does it just look like it does? Kubernetes HPA, a CI/CD pipeline I tore apart and rebuilt for speed, and a Gemini-powered evaluation engine with race-condition handling on concurrent AI calls.
Stack: Next.js · MongoDB · Kubernetes · GitHub Actions · Prometheus/Grafana · k6
Shadow Permission Analyzer — AWS IAM privilege-escalation detector (repo)
Models IAM identities as a graph instead of a flat policy list, because privilege escalation is a path problem, not a single-permission problem. BFS/DFS traversal finds multi-hop chains a manual audit would miss.
Stack: Python · Neo4j · FastAPI · boto3 · pytest
DocuFlow — event-driven invoice pipeline (repo)
OCR parsing and data extraction split into independently scalable workers, because a single fat script is not a pipeline. 40% faster processing through async orchestration.
Stack: Python · Celery · Redis · PostgreSQL
🚧 DriftGuard — infra drift detection engine (planning)
Compares live Docker/Kubernetes environments against Git-managed config and flags what quietly changed underneath you. Because "it worked when I deployed it" is not a monitoring strategy.
Planned stack: Python · FastAPI · Docker · Kubernetes · GitHub Actions
$ echo "always be shipping" >> /etc/motd