LON-UK · SYSTEM NOMINAL
AI Software Engineering & Distributed Systems · United Kingdom

Distributed AI &
Systems Architecture

Associate AI Software Engineer · Ex-Omnitech Business Ltd · MSc Computer Science

Designing production-grade multi-agent architectures, retrieval-augmented generation (RAG) pipelines, and high-concurrency Python/FastAPI backend infrastructure with real-time telemetry, deterministic guardrails, and cryptographic access control.

Career Experience Download CV (PDF) Interactive AI Lab

AI Engineer Profile & Specifications

UK INDUSTRIAL PLACEMENT VERIFIED

OPERATIONAL
Industrial Placement Omnitech Business Ltd
Production LLM & RAG Python · FastAPI · LangGraph
Event & Media Streaming WebSockets · AsyncIO
Automated Test Suite 198/198 Pytests Passing
LinkedIn Profile Direct Inquiry
Engineering Profile

Distributed AI Infrastructure & Systems Engineering

Backed by postgraduate training in Computer Science at Aston University (Birmingham, UK), my work focuses on translating large language models and quantitative algorithms into deterministic, fault-tolerant software systems.

I design and maintain end-to-end backend pipelines—spanning multi-agent retrieval architectures, bidirectional voice/WebSocket media streams, ACID-compliant telemetry ledgers, and hardware-authenticated remote management consoles.

65+

Production API Endpoints

160

Automated Pytest Assertions

15+

Industry Certifications

UK

Birmingham · Right to Work

Career Track Record

Engineering Experience

Industrial placement and enterprise systems leadership across the United Kingdom and global IT infrastructure.

Omnitech Business Ltd

ASSISTANT AI SOFTWARE ENGINEER UK PLACEMENT

London / Birmingham, United Kingdom · Full-Time Industrial Placement

June 2026 – Jan 2027

Engineered production-grade Generative AI microservices, deterministic retrieval-augmented generation (RAG) pipelines, and autonomous agent gateways across distributed cloud environments.

  • Production LLM Microservices: Developed asynchronous Python/FastAPI backend services integrating Anthropic Claude and Google Gemini APIs, delivering context-aware responses with <450ms P95 turnaround and automated token rate-limit retry logic.
  • Hybrid Vector Indexing & RAG: Architected hybrid dense-sparse retrieval systems (Chroma/Qdrant + BM25) utilizing Reciprocal Rank Fusion (RRF) and semantic caching, cutting query execution costs by 38% and eliminating hallucinated context.
  • Deterministic Agentic Workflows: Designed structured tool-calling agent pipelines with Pydantic v2 schemas and fail-closed validation guardrails, preventing hallucinated or malformed mutations on mission-critical database entities.
  • Continuous Evaluation & Benchmarking: Built continuous LLM-as-a-judge evaluation pipelines using the RAGAS framework across 100+ synthetic edge-case scenarios, boosting answer faithfulness from 78% to 94% prior to staging deployment.
  • Agile Delivery & Cloud Reliability: Containerized AI services with Docker, managed cloud deployments on AWS, and collaborated actively in bi-weekly Agile sprints, sprint demos, and technical RFC architectural reviews.
Python 3.12+ FastAPI Claude API Gemini API RAGAS Framework Hybrid RAG (BM25 + Chroma) Docker AWS

Wipro Technologies

TECHNICAL SUPPORT & WORKFLOW LEAD CUSTOMER EXCELLENCE AWARD

Pune, India · Enterprise Data Infrastructure & Technical Operations

April 2022 – Sept 2023

Supervised mission-critical enterprise storage and disaster recovery systems, directed shift queue triaging, and ensured strict SLA compliance for global clients in partnership with Dell Technologies.

  • Enterprise Systems Reliability: Maintained high-availability Dell Data Domain enterprise storage clusters and disaster recovery hardware/software, cutting system downtime by 25%.
  • Workflow & Queue Management: Directed queue allocation and SLA enforcement for 200+ monthly enterprise support tickets across a 15-member team, balancing workloads based on domain expertise.
  • Engineering Escalations & Recognition: Resolved 30+ tier-3 complex escalations in direct collaboration with Dell Technologies engineering, earning the Customer Excellence Award (recognized in top 5% of 200+ engineers).
Enterprise Storage Dell Data Domain RAID Diagnostics Linux Systems Workflow Allocation SLA Management
Technical Domains

Core Engineering Capabilities

Structured software architecture across inference pipelines, asynchronous backend services, and cloud infrastructure.

Retrieval-Augmented Generation & Agents

Engineering deterministic multi-agent workflows, semantic vector indexing, structured tool-use schemas, and multimodal LLM evaluation pipelines using Anthropic Claude and Google Gemini APIs.

Asynchronous Backend Services

Developing high-concurrency REST and WebSocket microservices in Python (FastAPI, AsyncIO), SQLite WAL persistence layers, headless browser worker pools (Playwright), and automated health watchdogs.

Cloud Infrastructure & Security

Implementing RFC 6238 HMAC-SHA1 TOTP authentication gates, request rate-limiting WAF middleware, encrypted tunnel orchestration, and automated CI/CD deployment workflows across AWS and Linux hosts.

Python 3.12 / 3.14 FastAPI & Starlette AsyncIO & WebSockets SQLite WAL (ACID) Anthropic Claude SDK Google Gemini API Multi-Agent RAG Playwright Headless Workers Twilio Media Streams AWS Cloud Infrastructure RFC 6238 TOTP Security Pytest Automated Verification
Live Interactive Bench

AI Architecture & System Design Lab

Interactive telemetry and architectural simulators demonstrating production trade-offs between RAG paradigms, multi-model consensus, and low-latency streaming.

SYSTEM DESIGN SIMULATOR v2.4

Retrieval Strategy & Ingestion Depth

Simulate architectural trade-offs across chunking strategies, vector index types (HNSW vs Flat), semantic caching, and re-ranking depth (Cohere/Cross-Encoder).

SIMULATED REAL-TIME METRICS
320ms
P95 TTFT
94.2%
Context Precision
1.8%
Hallucination Risk
$0.0018
Avg Query Cost
Active Architecture Profile: Combining HNSW dense vectors with keyword-exact BM25 via Reciprocal Rank Fusion ensures domain-specific acronyms and entities are never dropped, while keeping P95 latency sub-400ms.
Professional Qualifications

Verified Credentials

Completed technical certifications across cloud architecture, machine learning, and enterprise engineering.

Amazon Web Services

AWS Cloud Essentials & Architecture

VERIFIED CREDENTIAL

IBM

Applied Big Data Analytics

VERIFIED CREDENTIAL

Deloitte

Technology Consulting Program

VERIFIED CREDENTIAL

Accenture

Software Engineering Program

VERIFIED CREDENTIAL

Microsoft

Cloud & Systems Engineering

VERIFIED CREDENTIAL

DeepLearning.AI

Supervised Machine Learning Foundations

VERIFIED CREDENTIAL

Anthropic

Claude Code — Agentic Engineering

VERIFIED CREDENTIAL

Anthropic

Claude API & Production Workflows

VERIFIED CREDENTIAL
Production Case Studies

Selected Systems Architecture

Production software platforms engineered for multimodal communication, distributed telemetry, and automated deployment.

System Architecture · 01

Multimodal Communications & RAG Gateway

An asynchronous communications gateway built with Python, FastAPI, and Twilio Media Streams to ingest bidirectional voice and messaging payloads, execute vector-backed retrieval, and synchronize structured records with CRM and payment APIs.

  • Real-time speech-to-text transcription and document context retrieval
  • WebSocket media streaming with sub-second response latency
  • Deterministic webhook validation and CRM state synchronization
Python FastAPI Twilio WebSockets Vector RAG REST Webhooks
PIPELINE TOPOLOGY · WSS / REST
Twilio WSS PCM Audio / Msg FastAPI Core AsyncIO Router LLM + RAG Vector Context CRM State Engine · Webhook Dispatch · Audit Log SLA: < 420ms P95 Turnaround · 99.9% Uptime
CONTROL PLANE ARCHITECTURE
DXGI KVM Stream 60 FPS · WebAudio Worker Daemons Playwright + REST RFC 6238 TOTP Zero-Trust WAF SQLite ACID WAL Ledger & 65-Endpoint REST/WS Bus 160/160 Pytest Suite · Self-Healing Watchdog
System Architecture · 02

Enterprise Operations & Telemetry Platform

A centralized operational platform integrating 65 REST and WebSocket endpoints, an ACID-compliant SQLite WAL ledger, headless browser worker daemons, DirectX 11 DXGI remote workstation streaming, and hardware-backed RFC 6238 TOTP security.

  • Low-latency DirectX 11 desktop & WebSocket PCM audio streaming
  • Server-verified RFC 6238 HMAC-SHA1 TOTP + WAF rate-limiting shield
  • 160 automated unit & integration tests with continuous watchdog auditing
FastAPI SQLite WAL WebSockets Playwright RFC 6238 TOTP
System Architecture · 03

Automated Release & OAuth 2.0 Syndication

An end-to-end Python release automation pipeline that validates repository diffs, executes regression test suites, synchronizes encrypted state snapshots via TLS FTP, and publishes release notes via OAuth 2.0 REST APIs.

  • Automated build verification and atomic remote synchronization
  • OAuth 2.0 token lifecycle management and REST API syndication
Python OAuth 2.0 REST APIs CI/CD Automation
CONTINUOUS DEPLOYMENT FLOW
Git / Pytest 100% Pass Gate State Snapshot Vault Sync Edge Deploy TLS + OAuth 2.0
Direct Communication

Engineering & Architecture Inquiries

Available for Associate AI Software Engineer and Backend Systems roles across the United Kingdom. Reach out directly via email or LinkedIn to discuss technical roles or architecture consulting.

sanchitkaushal2002@gmail.com LinkedIn Download CV (PDF)