Srinath Doggala
AI Engineer building reliable AI systems, not just LLM demos.
I build production-oriented AI systems across LLM applications, agentic workflows, RAG, voice AI, and backend infrastructure using Python, FastAPI, React/Next.js, and PostgreSQL.
Srinath Doggala
AI Systems Engineer · Agentic Workflows
Building production-oriented AI systems involving LLMs, agents, RAG, voice AI, backend infrastructure, and full-stack product engineering.
Med. API Latency Reduction
Redis Caching + AsyncIO Pipelines
Measured on production endpoint request workloads at Sreeva AI
VoxPilot Automated Tests
Automated Unit & Integration Tests
Routing policies, voice streaming, RAG, circuit breakers & safety
Autonomous Research Swarm
Planner, Researcher, Verifier, Writer
Real-time web retrieval with automated claim verification
Resumes Evaluated
Continuous Load & ATS Scoring
Benchmarked parsing stability and vector embedding match consistency
Production-Oriented Case Studies
Engineering evidence across real-time voice streaming, multi-agent pipelines, developer tooling, and document intelligence.
VoxPilot AI: Real-Time Voice Agent Platform
Real-time voice AI infrastructure with WebSocket audio transport, adaptive model routing, selective RAG, and circuit breakers.
Real-time voice agents fail unpredictably in production due to upstream provider outages, latency jitter, unverified tool executions, and brittle session state.
A full-stack voice AI platform with WebSocket audio transport, STT/LLM/TTS provider abstractions, adaptive multi-agent routing, selective RAG, deterministic tool safety confirmation gates, provider failovers, circuit breakers, session replay, and token/cost telemetry.
Low-latency bidirectional streaming connecting browser audio to backend voice orchestrator.
Monitors upstream provider health and latency; automatically triggers fallback failovers upon errors.
Orchestrates intent classification with strict risk classification and approval gates before executing tools.
Contextual vector retrieval, per-request latency & cost telemetry, plus session replay for evaluation.
- 27 automated tests passing (routing, voice pipeline, RAG, circuit breakers, safety)
- Provider failure recovery with sub-second circuit breaker failovers
- Deterministic risk-classified tool execution schema
- Session replay and latency/cost telemetry dashboard
Frontend live on Vercel. 27/27 automated tests passing in CI. Backend requires cloud API credentials or local Docker runner.
ResearchGPT: Multi-Agent AI Research Assistant
Autonomous 4-agent research pipeline executing concurrent web retrieval, claim verification, and cited synthesis.
Single-prompt LLMs hallucinate citations and rely on stale training weights, producing unverified reports unfit for serious analysis.
An orchestrated multi-agent research pipeline where Planner breaks queries into targeted hypotheses, Researcher gathers web documents concurrently via Tavily, Verifier cross-references claims against raw sources, and Writer synthesizes cited Markdown/PDF briefs.
Deconstructs complex user queries into structured research sub-questions and search parameters.
Executes concurrent web queries via Tavily API and extracts full text content.
Cross-checks extracted claims against primary sources to systematically eliminate hallucinations.
Synthesizes multi-source data into formatted Markdown/PDF reports with verified inline citations.
- 4 synchronized agents with isolated JSON schema validation
- 15+ REST micro-endpoints supporting asynchronous document processing
- PostgreSQL audit trail storing reasoning steps and verified citations
ReviewGPT: AI-Powered Code Review Platform
GitHub repository scanner combining AST static analysis, security audits, and Gemini 2.5 Flash refactoring.
Developers push avoidable security flaws (hardcoded secrets, SQLi, XSS) and high-complexity functions that standard linters miss and code reviews overlook.
An automated repository analysis platform combining Python AST parsing for deterministic cyclomatic complexity and vulnerability patterns with Google Gemini 2.5 Flash for contextual line-level explanations and automated refactoring diffs.
Scans public GitHub repositories, retrieving multi-branch structures and source file AST representations.
Runs AST static checks for cyclomatic complexity, security risks (SQLi, XSS, exposed keys), and boundary bugs.
Prompts Google Gemini 2.5 Flash to generate contextual code explanations, quality metrics, and refactoring fixes.
Renders real-time repository health score, issue distribution, file complexity metrics, and AI assistant UI.
- Deterministic AST checks for SQLi, XSS, and hardcoded credential leaks
- Cyclomatic complexity metrics computed per function and file
- FastAPI serverless microservices deployed on Vercel with asynchronous GitHub ingestion
TalentLens AI: AI Recruiter & Career Intelligence Platform
Enterprise resume intelligence & ATS engine performing deep semantic document evaluation and recruiter simulations.
Job seekers lack objective feedback on resume formatting and keyword compatibility against modern semantic ATS engines.
An end-to-end recruitment intelligence system with asynchronous PDF/DOCX layout parsing, vector embedding matching, LLM qualitative scoring, and interactive interview question generation.
Asynchronous PDF/DOCX parsing using Python extractors into structured text chunks.
Generating dense vector embeddings for candidate experience vs job requirement matrices.
FastAPI REST service invoking OpenAI/Claude models for qualitative candidate evaluation.
Responsive telemetry interface rendering ATS match breakdown and interactive feedback.
- 500+ resumes benchmarked in continuous stress and parsing tests
- Strict Pydantic/Zod schemas ensuring quantitative score stability
- Sub-3s document parsing and vector matching pipeline
AI-Based Produce Recognition System
Edge-optimized computer vision system providing sub-200ms item classification and instant nutritional telemetry.
Edge produce identification requires lightweight model footprints while maintaining high accuracy across varied real-world lighting.
A MobileNetV2 transfer-learning classification pipeline integrated with Django REST APIs and CalorieNinjas macronutrient lookups.
Transfer learning neural network trained on 30 fresh produce categories.
High-throughput inference gateway receiving image payloads.
Automated API call retrieving real-time macronutrient breakdown for top predictions.
Interactive visual cards rendering macro charts and item confidence scores.
- Sub-200ms model inference runtime on web endpoints
- 30 produce categories classified with high accuracy
- Live integration with nutritional database APIs
Production Contributions
Hands-on engineering delivering asynchronous microservices, LLM workflows, and database caching.
Founding AI Full Stack Engineer Intern@ Sreeva AI
Engineering production-oriented AI features and backend microservices using Python, FastAPI, React/Next.js, PostgreSQL, and LLM APIs.
Core Engineering Responsibilities
- Built AI-powered full-stack features using Python, FastAPI, React/Next.js, PostgreSQL, and LLM APIs.
- Designed asynchronous backend workflows, validated REST APIs, and Redis-backed application infrastructure.
- Developed RAG and multi-step LLM workflows using LangChain/LangGraph with deterministic validation and failure handling.
- Reduced API response latency by ~35% through asynchronous processing and Redis caching.
Technical Stack & Tooling
Practical engineering capabilities across AI orchestration, asynchronous backends, modern web frontends, and core CS fundamentals.
AI & Autonomous Systems
(6)LangChain · LangGraph · Agentic Workflows
OpenAI · Claude · Gemini APIs · Fallbacks
Chunking · Vector Embeddings · Similarity Search
Pydantic Schemas · Tool Calling · Safety Gates
MobileNetV2 · TensorFlow · Transfer Learning
Session Replay · Cost Tracking · Latency Monitoring
Backend & Microservices
(6)AsyncIO · WebSockets · REST · Dependency Injection
AsyncIO · Event Loops · Concurrency · Pydantic
OpenAPI Specs · Validation · Error Handling
Bidirectional Audio Transport · Real-Time Events
Django REST Framework · ORM · Microservices
Redis Queues · Non-blocking IO · Workers
Frontend Engineering
(5)Hooks · State Architecture · Performance Profiling
Server/Client Components · SSR · API Routes
Strict Types · Generics · Interface Contracts
Custom Design Tokens · Responsive Layouts · Dark Mode
Spring Physics · Micro-Animations · Layout Transitions
Databases & Storage
(4)Schema Design · Indexing · ACID Transactions · SQL
Caching Layers · TTL · Pub/Sub · Session State
Similarity Search · Cosine Distance · Indexing
JSON Documents · Aggregations · Indexing
DevOps & Infrastructure
(5)Multi-stage Builds · Docker Compose · Container Isolation
Automated Test Suites · Linting · Build Pipelines
Serverless Deployments · Edge Runtime · Routing
Branching Workflows · Code Reviews · Semantic PRs
Shell Scripting · Process Management · CLI Tooling
Core Computer Science
(4)Arrays · Trees · Graphs · DP · Complexity
Design Patterns · Separation of Concerns · Clean Code
ACID · Concurrency Control · Query Planning
Processes · Threads · Memory Allocation · Virtualization
Engineering Proof & Reliability
Engineering claims backed by automated test suites, measurable performance benchmarks, public repositories, and fault-tolerant architectural designs.
27 Automated Tests Passing in VoxPilot
Comprehensive test harness validating provider routing policies, WebSocket audio transport, selective RAG retrieval, circuit breaker state transitions, and deterministic tool safety gates.
Pytest unit & integration test suite executed across mock audio and provider failure streams.
35% Median API Response Reduction
Measured optimization at Sreeva AI achieved by refactoring blocking synchronous routes to Python asyncio event loops paired with in-memory Redis caching.
Before/after benchmark comparing repeated query latencies on production endpoint traffic.
500+ Resumes Document Evaluation
Stress-tested TalentLens AI document ingestion across varied formatting, multi-column layouts, and ATS keyword extraction under continuous load conditions.
Automated batch processing test harness verifying parsing stability and schema adherence.
Deterministic Guardrails & Circuit Breakers
Decoupled tool safety confirmation schema preventing unverified actions, combined with health-aware provider fallback routing when LLM APIs degrade.
Simulated upstream failure injection with automatic sub-second fallback failover.
Dockerized Microservices & CI/CD
Multi-stage Docker builds paired with automated GitHub Actions workflows for continuous linting, testing, and deployment to Vercel edge networks.
Automated GitHub Actions runners on every commit with pull-request status gates.
100% Verifiable Open-Source Repositories
Complete transparency: VoxPilot, ResearchGPT, ReviewGPT, and TalentLens have public GitHub repositories with setup documentation and clean git history.
Inspectable source code, architecture diagrams, test fixtures, and environment templates.
Evidence Deep Dive: 27 Automated Tests Passing in VoxPilot
Category: Automated Testing • Method: Pytest unit & integration test suite executed across mock audio and provider failure streams.
Eliminating reliance on unverified statements by establishing measurable automated checks, structured schemas, and failure recovery.
Comprehensive test harness validating provider routing policies, WebSocket audio transport, selective RAG retrieval, circuit breaker state transitions, and deterministic tool safety gates.
Pytest unit & integration test suite executed across mock audio and provider failure streams.
End-to-End System Architecture
An interactive breakdown of how client interfaces, asynchronous APIs, agent swarms, databases, and safety layers connect in my projects.
AI Orchestration & Agent Swarms
Autonomous supervisors, selective RAG, and deterministic tool safety
Architectural Responsibilities:
- 01.Orchestrating specialized multi-agent roles (Planner, Researcher, Verifier, Writer)
- 02.Adaptive model routing with automatic health monitoring and failover
- 03.Decoupled risk classification and confirmation gates for tool executions
Academic Education
Formal foundations in Computer Science, Artificial Intelligence, Machine Learning, and Mathematics.
Chandigarh University
B.E. Computer Science Engineering (AI/ML)
- Specializing in Artificial Intelligence, Machine Learning, and Distributed Software Systems.
- Core Coursework: Data Structures & Algorithms, Deep Learning, Operating Systems, Database Management Systems, Computer Networks, Software Engineering.
- Active participant in technical coding clubs, hackathons, and AI research projects.
SR Junior College
Intermediate (MPC: Mathematics, Physics, Chemistry)
- Senior Secondary Board Examination: 97.5% aggregate percentage (Academic Distinction).
- Demonstrated analytical excellence in Advanced Mathematics, Physics, and Chemistry.
Verified Certifications & Credentials
Validated domain coursework in Google Cloud AI, Generative AI, Machine Learning, and SQL
Innovating with Google Cloud AI
Vertex AI · Google Cloud ML · Generative AI Solutions
Generative AI: Introduction and Applications
Foundation Models · LLM Prompting · Enterprise GenAI
Introduction to Machine Learning
Supervised Learning · Regression & Classification · Neural Nets
The Structured Query Language (SQL)
Complex Joins · Subqueries · Database Schema Optimization
Looking for My Next Engineering Opportunity
I am currently interviewing for AI Engineer, Applied AI, AI/ML Engineer, and AI Full-Stack roles.
Founding AI Full Stack Engineer Intern @ Sreeva AI. Ready for immediate technical discussions, code walk-throughs, and system design interviews.
One-page verified summary covering system architectures, automated testing benchmarks, and professional experience.
Download Resume (PDF)Direct Message or Interview Inquiry
Submitting directly opens an email draft addressed to doggalasrinath@gmail.com