Nischay Gupta
Solution Designer · AI/Platform Engineer · Solutions Architect
- [email protected]
- nischaygupta.dev
- linkedin.com/in/itsnischayyy
- github.com/itsnischayyy
- Surat, Gujarat, India
Summary
- 3+ years shipping production systems, now owning solution design across concurrent enterprise client platforms — from problem framing and architecture through prototype, hardening, production rollout and client UAT sign-off. Previously Lead Developer for a full product team.
- Builds applied AI and data platforms in Python: production hybrid RAG (vector + full-text retrieval, rank fusion, cross-encoder reranking, grounded citations), speech-to-text with speaker diarization, entity canonicalisation and self-improving taxonomies, and local GPU inference for data that cannot leave the network.
- Architects backends in Java/Spring Boot, C#/.NET Core, Node.js/TypeScript and Python — layered, event-driven design with queue-based ingestion, realtime sockets and multi-provider normalisation — deployed across AWS, Azure and GCP.
- Security and governance by default: Microsoft Entra SSO (OIDC/PKCE), OAuth2/JWT, RBAC with database row-level security, AES-256-GCM client-side encryption, TOTP 2FA and append-only audit trails.
Experience
Also embedded with Connexink as Solutions Architect & Senior Engineer for their outsourced client accounts.
- Architected a field feedback intelligence platform for a global consumer-health client: unstructured text and multi-speaker Hindi/English voice recordings become RBAC-scoped business intelligence for a five-tier sales hierarchy, with all inference on local GPU.
- Built a hybrid RAG assistant behind a question-type router — statistical questions hit TimescaleDB rollups at zero LLM cost; qualitative ones run vector + full-text retrieval with RRF, cross-encoder rerank and grounded citations, capped at two LLM calls per answer.
- Enforced RBAC in three independent layers: API scope dependencies, Postgres row-level security via per-session GUC, and scope filters injected into the vector query.
- Closed a self-improving vocabulary loop (HDBSCAN drift detection, schema-constrained LLM induction, semantic guardrails) that took the review queue from 22 near-duplicate proposals to effectively zero.
- Diagnosed a native-library load-order fault silently killing workers and an IPv6 resolution issue adding 2s to every model call — 149x faster vector queries.
- Architected telematics ingestion for an EV fleet platform (webhooks → SQS → unified schema, DLQ auto-retry, per-provider health) — shipped with 148/148 backend and 106 client UAT cases passing.
- Delivered an enterprise event & budget lifecycle portal for a pharmaceutical client on ASP.NET Core 8 + Dapper + SQL Server with Azure AD SSO.
- Built a data intelligence & network visualisation platform turning unstructured social data into embeddings, semantic clusters and an interactive influence graph.
- Architected a User Management System from scratch on Spring Boot serving millions of API requests, with fine-grained role & permission authorization across hundreds of endpoints.
- Designed and shipped two production React Native apps (User and Agent) with real-time chat and voice/video over WebRTC on custom signalling, scaling to 10,000+ concurrent users.
- Ran infrastructure on AWS and GCP over PostgreSQL, MongoDB and Redis at a 99.9% uptime target; Jenkins/Docker pipelines and Prometheus monitoring cut manual deployment effort by 60%.
- Built an AI audio analytics platform in C#/.NET analysing millions of support calls at ~2x previous throughput, with an LLM evaluation layer scoring agents at 97% agreement with human reviewers.
- Delivered a facial recognition engine at 98% accuracy, matching 5,000–10,000 images in under 10 seconds via an optimised OpenCV + CUDA embedding pipeline.
- Shipped an AI form-processing service classifying and routing 100,000+ submissions with automated email workflows (Python + NestJS).
- Automated 200-page report generation in 20 seconds with Puppeteer, cutting report storage by 80%.
Key projects
Turns code-mixed voice notes from the field into governed, queryable business intelligence — with every model running on local GPU.
0 LLM calls for statistical questions · ≤2 LLM calls per qualitative answer · 22 → 0 near-duplicate taxonomy proposals · 149x faster vector queries after IPv6 fix · 3 independent RBAC layers
Python, FastAPI, Celery, SQLAlchemy 2.0 (async), PostgreSQL, TimescaleDB, Qdrant, Redis, MinIO, faster-whisper, pyannote, BGE-M3, Cross-encoder reranker, vLLM, Ollama (Qwen), HDBSCAN, Microsoft Entra ID, Docker
One unified telemetry stream from many vendors, powering a platform that runs an electric delivery fleet end to end.
148/148 backend tests passing · 106 client UAT cases passed · 0 deploys needed to add a provider
Java Spring Boot, Node.js, Express, MongoDB, AWS SQS, Redis, Socket.IO, BullMQ, Next.js, React Native, Expo, TypeScript, TypeORM, PostgreSQL, S3, Docker, ECR, GitHub Actions
Medical-education events tracked from money down to evidence of delivery.
ASP.NET Core 8 MVC, C#, Dapper, SQL Server, Microsoft Entra ID, Azure Blob Storage, Quartz.NET, Serilog, jQuery, Bootstrap 5
Raw social data → embeddings → communities, explored as an interactive influence graph.
Python, Vector embeddings, Clustering, Graph visualisation, Multi-source data integration
Millions of support calls scored by LLM, and a face matcher that searches 10,000 images in under 10 seconds.
97% agreement with human reviewers · 98% face recognition accuracy · <10s to match 5,000–10,000 images · ~2x call-analysis throughput
C#, .NET Core, Python, OpenAI APIs, OpenCV, CUDA, SQL Server
Centralised identity and access for a marketplace, serving millions of API requests.
Millions API requests served · 99.9% uptime target · -60% manual deployment effort
Java, Spring Boot, Spring Security, Spring Data JPA, JWT, PostgreSQL, Redis, AWS, GCP, Docker, Jenkins, Prometheus
Two live Play Store apps with chat and voice/video calling over WebRTC, scaled to 10,000+ concurrent users.
10,000+ concurrent users · 2 apps live on Google Play
React Native, WebRTC, Socket.IO, Java Spring Boot, PostgreSQL, MongoDB, Redis, AWS
Skills
- Languages
- Python, Java, C#, TypeScript, JavaScript, SQL, C++
- AI & ML
- Hybrid RAG, LLM evaluation & guardrails, LLM-as-judge, Embeddings (BGE-M3), Cross-encoder reranking, ASR + speaker diarization, Reciprocal Rank Fusion, Grounded citations, Prompt engineering, Structured / schema-constrained outputs, Function calling, Tool-calling agents, MCP servers, HDBSCAN clustering, faster-whisper, pyannote, OpenCV, Face recognition
- AI Serving
- vLLM, Ollama (Qwen), OpenAI APIs, Local GPU inference, Provider-abstracted model layer, GPU profile tuning, CUDA
- Backend
- Spring Boot, FastAPI, .NET Core / ASP.NET Core, Node.js, Microservices, Spring MVC, Spring Security, Spring Data JPA, Hibernate, Maven, Gradle, Celery, SQLAlchemy 2.0 (async), Dapper, Express, NestJS, TypeORM, REST, WebSockets, Socket.IO, WebRTC, Queue-driven ingestion
- Architecture
- Event-driven, Microservices, Clean / N-Tier, DDD, MVC, SOLID, API Gateway, Design patterns (DI, Repository, Factory, Strategy, Adapter)
- Messaging
- AWS SQS, Dead Letter Queues, Celery, BullMQ, Socket.IO, Webhooks
- Data
- PostgreSQL, TimescaleDB, Qdrant, Redis, Row-level security, Full-text search, SQL Server, MongoDB, MySQL, MinIO, Alembic, Schema design, Query optimisation
- Cloud & DevOps
- AWS, Azure, Docker, Kubernetes, EC2, S3, ECR, Secrets Manager, Entra ID, Blob Storage, GCP, GitHub Actions, Jenkins, Prometheus, Serilog
- Security
- Microsoft Entra SSO (OIDC/PKCE), OAuth2 / JWT, RBAC & row-level security, AES-256-GCM, Spring Security, PBKDF2, TOTP 2FA, Append-only audit logging, PII handling
- Frontend & Mobile
- React, Next.js, React Native, Expo, Angular, Redux, Tailwind, Razor, jQuery, deck.gl, Mapbox
- Practices
- Agile / Scrum, Code review, CI/CD, TDD