Hello
SOMANI ABDULLA
SURYA TEJA
Full Stack & Applied AI Engineer
1.9+ Years High-Scale Production Engineering @ Codegnan · Dec 2024 – Sep 2026.
FullStack&AppliedAIEngineershippingenterprisewebplatforms,scalableAPIs,andreal-timevoiceAI.Owneddeliveryacross10productionrepositoriesserving5,000+users(246PRs,963commits).SpecializedinTypeScript,Python(FastAPI),Node.js,Next.js,Docker,AWSandVitest.
Scaling systems at the
edge of intelligence
and cloud resilience.
Specializing in low-latency voice synthesis pipelines with LiveKit WebRTC, high-throughput MongoDB multi-tenant microservices, and end-to-end production ownership.
High-Throughput Microservices
Eliminated MongoDB query bottlenecks with compound indexing, driving down query latency by 68% across high-concurrency workloads.
Applied Voice AI Pipelines
Full-duplex conversational agents with DeepSeek R1 and OpenAI streaming through LiveKit WebRTC for human-parity responsiveness.
End-to-End Ownership
From database schema to production Linux Nginx reverse proxies with automated zero-downtime CI/CD workflows.
Every breakthrough begins in
absolute stillness.
Before writing a single line of code, I break complex distributed challenges down to first principles. Whether optimizing MongoDB index cardinality or architecting full-duplex WebRTC pipelines, clarity precedes speed.
Building systems that scale
without breaking.
Proven track record across 1.9+ years at Codegnan IT Solutions: managing 246 PRs, 963 commits, and deploying multi-tenant production architectures that reliably serve thousands daily.
Selected Works
Enterprise EdTech and real-time voice screening platforms shipped to live production users.
Enterprise LMS Platform
Microservices learning ecosystem scaled to 5,000+ active users
AI Mock Interview with Real-Time Voice
Ultra-low latency conversational AI screening rooms
Placement & Candidate Screening Portal
High-throughput recruitment workflow and eligibility engine
Payments & Billing Automation
Idempotent payment reconciliation with zero duplicate charges
Architected Platforms
Enterprise LMS Platform
Microservices learning ecosystem scaled to 5,000+ active users
- ◆Scaled microservices LMS to 5,000+ users across 156 REST endpoints and ~24 data models.
- ◆Authored 97% of backend commits, architecting course management and analytics.
- ◆Optimized high-traffic dashboards to maintain sub-100ms response times.
AI Mock Interview with Real-Time Voice
Ultra-low latency conversational AI screening rooms
- ◆Achieved ~650ms conversational glass-to-glass latency using LiveKit WebRTC.
- ◆Self-hosted STT, LLM and TTS pipelines behind asynchronous FastAPI streaming endpoints.
- ◆Engineered adaptive bitrate fallbacks ensuring zero conversational interruptions.
Placement & Candidate Screening Portal
High-throughput recruitment workflow and eligibility engine
- ◆Built placement and screening portal connecting 500+ corporate recruiters.
- ◆Implemented dynamic eligibility reach-meter calculating applicant qualification scores.
- ◆Secured 40+ endpoints with 5-tier role-based access control and httpOnly JWT tokens.
Payments & Billing Automation
Idempotent payment reconciliation with zero duplicate charges
- ◆Engineered Razorpay checkout flows (full, token, balance) with SHA-256 signature verification.
- ◆Guaranteed idempotent transaction handling with zero duplicate charges over 1,000+ payments.
- ◆Automated Zoho invoice generation, saving 15+ manual finance hours weekly.
Autonomous Career Intelligence Engine
Modular multi-tier ATS reverse-engineering automation pipeline
- ◆Automated job discovery across 5 enterprise portals using an Adapter Pattern engine.
- ◆Categorized opportunities into APPLY, WATCH, NEAR, and SKIP tiers based on ATS schema match.
- ◆Integrated two-way Excel sync and real-time HTML analytics dashboard.
Capabilities & Tooling
Frontend Architecture01
Backend & Systems02
Applied AI03
Data & DevOps04
Quality & Testing05
Production Experience
MERN Stack Developer (Full Stack & AI Engineer)
- 01.Lead architect across 5 live production platforms, authoring 246 PRs and 963 commits.
- 02.Achieved 200x query speedup with targeted compound indexes and aggregation pipeline optimization.
- 03.Slashed core API latency from 1,000ms+ down to 40-50ms under peak concurrent learner traffic.
- 04.Reduced server CPU by 40% with direct AWS S3 presigned URL uploads and HLS adaptive streaming.
- 05.Maintained 80%+ test coverage with Vitest, passing SonarQube quality gates on all releases.
- 06.Secured 40+ endpoints with 5-tier RBAC, JWT access tokens, and automatic refresh token rotation.
Proven track record at Codegnan IT Solutions. Scaling distributed systems, voice AI agents, and enterprise architectures with zero compromise.
An engineer who owns
the complete outcome.
I am a Full Stack & Applied AI Engineer who takes complete ownership: from database schema architecture and real-time WebSocket state synchronization, to fluid micro-interactions and Linux production deployment.
At Codegnan IT Solutions (Dec 2024 – Sep 2026), I architected and maintained core multi-tenant microservices serving over 15,000+ daily production requests. By optimizing MongoDB compound indexing and query execution paths, we eliminated blocking bottlenecks and accelerated read latency by 68%.
In the Applied AI domain, I developed real-time low-latency voice synthesis pipelines with LiveKit WebRTC, streaming DeepSeek R1 / OpenAI responses at sub-500ms response windows for seamless conversational intelligence.
B.Tech in Computer Science & Engineering
DRK Institute of Science and Technology
Diploma in Computer Engineering
Government Polytechnic Masab Tank
Cyber Security / VAPT Certification
Berry9 IT Services
Vulnerability Assessment & Penetration Testing certification