Open to software engineering roles
Abhinav Ranjan Sulabh NIT Patna · CSE · 2026
Backend systems. Cloud-native delivery. Evidence over adjectives.
Software engineer with Oracle Cloud Native SBC experience, a 9.72/10 CGPA, and documented work across Java backend systems, Kubernetes automation, forecasting pipelines, and privacy-preserving AI.
00 / Shortlist signal
What an HR screen sees in 20 seconds.
What engineers can verify in twenty minutes.
The page is intentionally layered: role fit and outcomes first, implementation decisions and trade-offs one click deeper.
Production exposure
Oracle Cloud Native SBC work across backend tooling, secure configuration, autoscaling, and observability.
XML · KEDA · Kubernetes · Prometheus · ThanosSystems depth
Projects go past CRUD into queues, caching, persistence, DNS resolution, cryptographic validation, and failure behavior.
Concurrency · JDBC · TTL cache · DNSSECMeasured engineering
Metrics are tied to their context: production internship impact, fixed out-of-sample evaluation, local benchmark, or simulation.
Impact without inflated claimsClear communication
Architecture notes, setup paths, QA outputs, runbooks, and technical reports make the work reviewable before an interview.
Documentation as engineering evidence01 / Selected engineering work
Four projects. Four different kinds of technical judgment.
Every project below is backed by a detailed README in the source material.
Applied ML · Energy markets
European Power Fair Value Forecasting
52%
lower OOS MAE
A reproducible DE-LU day-ahead forecasting pipeline that treats data quality, temporal validation, uncertainty, and reporting as first-class engineering concerns.
Public data→QA + features→Model validation→Fair value
Open engineering brief
Evaluation Fixed 2025 out-of-sample MAE improved from 32.83 to 15.75 EUR/MWh against a seasonal-naive baseline.
Pipeline Raw JSON caching, UTC/DST-safe cleaning, coded QA, feature engineering, blocked validation, and P10/P50/P90 reporting.
AI control LLM-proposed QA rules are filtered, logged, and backed by deterministic checks; code retains authority over the pipeline.
Java · Backend systems
Core Banking Engine
10K+
simulated daily flows
A modular Java banking engine covering account lifecycles, secure APIs, queue-backed transactions, pluggable persistence, migrations, and cache behavior.
JWT API→Domain→Queue→JDBC + cache
Open engineering brief
Design Layered architecture around a Bank aggregate, three account types, four transaction flows, repositories, observers, and factories.
Runtime Token-secured HTTP endpoints, asynchronous execution using a blocking queue and executor service, health and metrics surfaces.
Benchmark Two TTL cache scopes reduced repeated-read latency by 35% in a controlled local comparison.
Federated learning · Security
FED_IIDS
91.03%
simulated accuracy
A privacy-preserving intrusion detection framework where heterogeneous clients train collaboratively without sharing raw network traffic.
HospitalFactory→DP-SGD→FedAvg→Global IDS
Open engineering brief
Privacy Client-side gradient clipping and Gaussian noise reduce data leakage risk while raw traffic remains local.
Data engineering Correlation filtering and LightGBM ranking reduce 74 features to the strongest 30 inputs.
Result The simulated federated setup reaches 91.03% accuracy, within 1.9% of the centralized baseline.
Java · Networking systems
DNS Resolution Engine
10
DNS record types
A runnable Java 17 simulation that maps DNS hierarchy, resolution, caching, routing, and security concepts to concrete classes and execution traces.
Root→TLD→Authoritative→TTL cache
Open engineering brief
Resolution Root, TLD, and authoritative hierarchy with recursive and iterative lookup plus cache hit/miss tracking.
Reliability Round robin, anycast failover, TTL trade-offs, negative caching, and split-horizon behavior across eight runnable demos.
Security Real RSA-2048 signing and verification model a KSK-to-ZSK-to-RRSIG DNSSEC chain of trust.
02 / Production experience
Where the work met real users, constraints, and operations.
Outcome-oriented summaries aligned with the single resume included in this portfolio.
Oracle Cloud Native SBC
Cloud Infrastructure and DevOps Intern
- 40%Manual configuration eliminated through recursive XML parsing, path generation, and schema validation.
- 25%Idle resource consumption reduced through KEDA-driven Kubernetes autoscaling with Prometheus and Thanos signals.
- Secure by defaultInput sanitization, defensive error handling, and modular certificate provisioning strengthened multi-environment CLI workflows.
IIT Ropar Student support platform
Software Development Intern
- 300+Students served across five or more departments through recommendation and support workflows.
- 15% / 20%Higher engagement and lower drop-off after personalized guidance and query-support improvements.
- 30%Faster ticket-resolution workflows through REST APIs, JWT authentication, RBAC, analytics, and user-management modules.
03 / Engineering range
Enough breadth to collaborate. Enough depth to be questioned.
Languages
Java, Python, JavaScript, SQL, C, Bash, XML
Backend + data
REST APIs, Node.js, Express, FastAPI, JDBC, MySQL, MongoDB, JWT, RBAC, multithreading
Platform
Kubernetes, OpenShift, Docker, Helm, KEDA, Prometheus, Thanos, Linux, CI/CD
AI + analytics
Scikit-learn, Pandas, NumPy, forecasting, Flower, DP-SGD, model validation, automated reporting
04 / Credentials
Academic consistency and competitive foundations.
National Institute of Technology Patna
B.Tech in Computer Science and Engineering
9.72 / 10
CGPA · Aug 2022 - May 2026
LeetCode KnightTop 5.5% globally · peak rating 1864↗
CodeChefRating 1736 · rank 87 in Challenge 197↗
JPMorgan Chase Code for Good 2025NGO platform for patient tracking, care allocation, attendance, and SMS alerts◆
Open source + scholarshipUSACO Guide Java contributor · NTSE Scholar◆
05 / Resume
One resume. One coherent engineering narrative.
A focused, one-page software engineering resume aligned with the experience and projects shown here.
Available for SDE opportunities