╔══════════════════════════════════════════════════════════════╗
║ ║
║ > whoami ║
║ Vishad Mandal — Senior Software & AI Engineer ║
║ Building distributed systems & shipping AI ║
║ products to production since 2020 ║
║ ║
╚══════════════════════════════════════════════════════════════╝
class Engineer:
name = "Vishad Mandal"
role = "Senior Software & AI Engineer"
experience = "6+ years"
focus = ["Distributed Systems", "Backend Engineering", "AI Products"]
currently = "Building reliability into systems that can't afford to fail"
highlights = {
"Elasticsearch" : "Stabilized 9.5TB cluster → crashes 4-5/mo → <1/mo",
"Query Latency" : "8s → 1.2s (~85% improvement)",
"VKYC Platform" : "Recording failure 12% → 1.8% (Airtel PB + Union Bank)",
"AI Products" : "3 independent systems shipped to AWS production",
"Airflow DAGs" : "50+ owned across distributed on-prem worker nodes",
}
principle = "Production is the only environment that matters."Backend & API
AI & Machine Learning
Data Engineering
Databases
Cloud & Infrastructure
Observability & DevOps
All production projects are under NDA. Sanitized versions and architecture writeups coming soon.
┌─────────────────────────────────────────────────────────────────┐
│ 🟢 agentpolicy [ OPEN SOURCE ] │
│ Runtime RBAC + prompt injection defense for LLM agents │
│ Defines what tools an agent can call, what data it can │
│ access, and what happens when the LLM gets compromised. │
│ Permission ceiling propagation across delegation chains. │
│ Stack: Python · LangGraph · YAML · Pydantic │
│ → github.com/vishadmandal/agentpolicy │
├─────────────────────────────────────────────────────────────────┤
│ 🔴 EngiCore [ PRIVATE ] │
│ AI document intelligence — RFQ compliance mapping │
│ Hours of manual clause cross-referencing → minutes │
│ Budget-aware paragraph packing fills 400k context window │
│ Dual-retrieval: spec embeddings + feedback memory │
│ Stack: FastAPI · LanceDB · GPT · PyMuPDF · OCR · React │
├─────────────────────────────────────────────────────────────────┤
│ 🔴 Quantum AI [ PRIVATE ] │
│ Multi-tenant AI social publishing platform │
│ Hybrid AI: visual analysis + real-time trend fusion │
│ Handles all Meta/Instagram OAuth edge cases (3 paths) │
│ Stack: FastAPI · Gemini · Perplexity Sonar · Meta API │
├─────────────────────────────────────────────────────────────────┤
│ 🔴 Reelmaker [ PRIVATE ] │
│ AI-driven short-form video renderer from YAML templates │
│ Vision LLM analyzes clips → LLM maps to scene slots │
│ Cinematic logo finale via per-frame PIL/NumPy compositing │
│ Stack: Python · MoviePy · Ollama · OpenAI · Pydantic │
├─────────────────────────────────────────────────────────────────┤
│ 🔴 AI_VID [ PRIVATE ] │
│ Local-first AI video transcription + captioning platform │
│ Zero cloud dependency — all inference via Ollama │
│ faster-whisper + LLaVA + Celery on 6-service Compose │
│ Stack: FastAPI · Celery · Redis · MinIO · Ollama │
├─────────────────────────────────────────────────────────────────┤
│ 🟡 stripdown [ OPEN SOURCE — COMING SOON ] │
│ npm package surgeon — strips unused code at module level │
│ Babel AST analysis → synthetic wrapper via Node resolver │
│ No bundler, no build step, no monkey-patching required │
│ Stack: Node.js · Babel · Commander │
│ → github.com/vishadmandal/stripdown │
├─────────────────────────────────────────────────────────────────┤
│ 🔴 Myndfi [ PERSONAL — WIP ] │
│ AI-powered expense tracker + financial psychology engine │
│ Prophet forecasts spending trends + future projections │
│ Goal-aware onboarding (house, savings, retirement etc) │
│ LLM analyzes full spend history → money psychology report │
│ Web + mobile · Backend + UI built · Pending deployment │
│ Stack: FastAPI · Prophet · Airflow · Ollama · React │
├─────────────────────────────────────────────────────────────────┤
│ 🔵 TataPlay Workforce Forecasting [ PRODUCTION ] │
│ Prophet-based demand forecasting — 95-98% accuracy │
│ Reverse-engineered undocumented legacy DSS from scratch │
│ Stack: Python · Meta Prophet · Apache Airflow │
└─────────────────────────────────────────────────────────────────┘
[2020-07] Took ownership of live banking VKYC platform mid-college
[2021-03] Running 2 enterprise banking clients in production (Airtel PB + UBI)
[2022-06] Graduated B.Tech — never stopped shipping
[2023-01] Deployed first AI-powered product to AWS production
[2024-07] Stabilized 9.5TB Elasticsearch cluster — 4-5 crashes/mo → near zero
[2026-04] Currently: Senior Engineer @ ValueLabs · Open to new challenges
PING interests.local
reply from: E-Sports & Sim Racing — always optimizing lap times
reply from: Photography — finding signal in visual noise
reply from: Playing Drums — rhythm is just distributed timing
reply from: Go-Karting — latency matters everywhere
// Currently open to Senior Backend / SRE / AI Engineering roles
// Hyderabad, India · Open to hybrid & remote
