ABV-IIITM Gwalior · Information Technology · Class of '26
Software engineer who builds scalable backend systems and distributed infrastructure — not just features.
My work spans cloud-native platforms, production-grade RAG systems, and merged contributions to CNCF's core observability and orchestration ecosystem.
🌐 CNCF Contributor Jaeger · OpenTelemetry · Kubernetes (PRs merged into K8s v1.36, OTel main)
🏢 SDE Intern AlgoUniversity (YC S21) — Kubernetes · Redis · Docker · AWS
🎓 B.Tech Thesis UAV obstacle-aware path planning via Deep RL (D3QN / PPO)
🏆 McKinsey Fellow Forward Leadership Program
⚡ CP LeetCode Knight 1827 · 1200+ solved · 88% acceptance · 365-day streak
|
|
|
Merged code running in production systems across thousands of cloud-native deployments globally.
| Project | PR | Impact |
|---|---|---|
| OpenTelemetry | #46506 |
AST constant folding for pure OTTL functions — parse-time compiler optimization, eliminates redundant per-datapoint evaluation in collector hot path. Accepted by OTel maintainers. |
| Kubernetes | #137398 |
Goroutine hot-loop fix merged into K8s v1.36 in client-go StartEventWatcher. Reviewed and merged by senior K8s maintainers. |
| Kubernetes | #137999 |
Fixed kubelet cgroup resize nil-pointer panic. Merged by K8s maintainers. |
| Jaeger | multiple | MaxTraceSize safeguard (OOM prevention on large trace queries) · API v3 limit/attribute filtering · OTTL delete matching values — all with full unit test coverage, merged to main. |
GSoC 2026 · Proposed across 3 CNCF orgs: Jaeger AI Skills Framework · Kubeflow OTel Integration · Meshery AI Adapter — spanning Go compiler internals, Python SDK observability, and LLM agent interfaces.
Production-grade RAG system over the GitLab Handbook with dual-path retrieval, SSE streaming, and sub-200 ms latency
| Layer | Implementation |
|---|---|
| Retrieval | Dual-path FAISS + BM25 with Reciprocal Rank Fusion (k=60) |
| Generation | Llama 3.3 70B on Groq · SSE streaming · 0–100 confidence scoring |
| Chunking | Heading-aware chunker (350–500 tokens) · 10-rule generation protocol |
| Latency | Sub-200 ms retrieval · zero external API calls in search path |
| Infra | Single Docker multi-stage image (Node 20 + Python 3.9) · admin analytics dashboard · live re-ingestion endpoint |
Concurrent endpoint health-checking engine with retry logic, timeout propagation, and bounded-latency execution
- Built in Go with native concurrency primitives for parallel endpoint monitoring
- Retry logic + timeout propagation ensuring bounded-latency execution under load
- Designed for reliable health checking across distributed service surfaces
- Clean, production-ready monitoring engine with zero external dependencies
Full-stack competitive programming platform — Kubernetes-native, Redis-streamed, Docker-sandboxed
- Docker-isolated compiler microservice with cgroups + SECCOMP sandboxing
- Redis Streams decoupled eval pipeline · JWT + OAuth2 + RBAC auth
- Live contests · leaderboards · submission analytics · rating system
📌 LeetCode Knight · Rating 1827 · 1200+ solved · 88% acceptance · 365-day consistency badge
📌 Flipkart GRiD Round 3 reached (cleared 2 coding rounds)
📌 AlgoU Camp Top 0.5% nationally
UAV-Assisted Obstacle-Aware Data Collection for Resource-Constrained IoT Networks
Supervisor: Dr. Ankur Jaiswal, ABV-IIITM Gwalior
Python · NumPy · D3QN · PPO · Visibility Graph
- Modelled 800×600 m IoT deployment — 20 nodes, 3 cuboid obstacles (heights 128–148 m > UAV altitude 100 m)
- Greedy dominating-set rendezvous selection achieves 65% node compression (20 → 7 stops, 100% coverage)
- Visibility-graph routing produces collision-free paths (~2566 m, 9.2% battery)
- Formulating D3QN agent for adaptive obstacle-aware path optimisation — state encodes UAV pose, residual energy, obstacle proximity, RP visit vector; reward jointly optimises data-priority collection, collision penalty, and energy cost
- Exploring PPO as continuous-action extension
Published on Medium with 500+ reads per article, covering complex architectures with visual intuition and real-world analogies.
Topics: System Design · Advanced DSA · SOLID Principles · Graph Algorithms · Load Balancers · Caching Strategies · Distributed Systems
"If it scales — it succeeds. Build with impact, not just intent."