Kamal Poshala
Full-Stack Software EngineerApplied AI & Distributed Systems
Specializing in building scalable web applications, robust backend architectures, and integrating applied machine learning models. I bridge the gap between complex AI research and production-ready systems, turning innovative concepts into reliable, high-performance software.
Professional Summary
/System Architecture & AI Overview
I am a Full-Stack Software Engineer passionate about engineering backend systems, seamless frontend experiences, and intelligent AI pipelines designed to operate at scale.
My expertise lies extensively in building multi-service ecosystems, scalable full-stack applications (React/Node.js), and transforming academic machine learning concepts (NLP/Transformers) into robust, fault-tolerant environments that remain highly available under heavy workloads.
Featured Projects
/Recent implementations in Distributed Systems & Scale.
Architected real-time collaborative platform supporting concurrent document editing, live chat, and presence tracking. Designed room-based event-driven communication with modular socket handlers. Implemented JWT-secured REST and WebSocket auth with RBAC. Deployed backend on Koyeb and frontend on Vercel.
Designed distributed approval workflow with role-based access control and secure JWT authentication. Implemented 15+ RESTful endpoints handling 1,000+ transaction events with centralized logging. Automated email remindes using cron, reducing approval turnaround by 40%.
Developed TF-IDF based search engine indexing 10K+ academic documents with relevance scoring. Built OCR and NLP ingestion pipeline transforming 2K+ PDFs into structured datasets. Implemented cosine similarity ranking and reduced query latency by 45%.
AI / LangChain
AI Book Summarizer & Vector Store
Engineered an automated summarization tool integrating ChromaDB and HuggingFace embeddings for precise vector storage and retrieval of long-context book texts via semantic search.
AI / Audio Processing
Automated Video Dubbing Engine (Deepdub)
Built a multi-language video dubbing pipeline replacing traditional translation dependencies with deep-translator for stable lip-sync TTS conversion and automated audio track replacement.
Professional Experience
/Enterprise Observation & Engineering
System Uptime
99.8%
Throughput
+35%
Latency
-25%
Incidents Resolved
20+
Software Engineer Intern
Cotiviti India // Hyderabad, India
- >Delivered 12+ production-grade RESTful services processing 5K+ daily transactions in distributed backend architecture, increasing automation efficiency by 30%.
- >Eliminated 40% of manual operational workflows by designing microservice-based internal platforms adopted across 3 engineering teams.
- >Increased backend throughput by 35% via asynchronous processing models, database indexing strategies, and aggregation query tuning.
- >Resolved 20+ high-priority production incidents through systematic debugging and root cause analysis, improving service availability to 99.8%.
- >Partnered with DevOps teams to streamline CI/CD deployment workflows, reducing release cycle time and improving delivery reliability.
Software Engineer Intern
Persistent Systems // Hyderabad, India
- >Developed performance-focused Java backend modules reducing execution time by 25% through memory optimization and algorithmic refinement.
- >Implemented scalable data structure-driven solutions (HashMaps, Heaps, Graph traversal) for compute-intensive engineering tasks.
- >Enhanced multi-threaded application stability by improving thread synchronization and CPU resource utilization.
Education & Certifications
/Academic Milestones and Cloud Native Credentials
Academic Training
Master of Science in Computer Science
University of Oklahoma • Norman, OK
- ▸ Specialization: Artificial Intelligence & Distributed Systems
Bachelor of Technology in Information Technology
CVR College of Engineering • Hyderabad, India
- ▸ Final CGPA: 9.08/10.0
- ▸ Distinctive Academic Achievement
Academic Highlights
/Advanced Implementations and Research Labs
Multilingual NLP Engine
Designed and built an academic proof-of-concept multilingual analysis engine leveraging transformer-based models (mBERT) for polarization detection and sentiment analysis across 22 complex languages without performance degradation.
Vector Extraction
High-dimensional representations
Cross-Lingual Zero-Shot
Transformer-based NLP
Technical Toolkit
/Languages, Frameworks, and Infrastructure
Layer 4: Systems & DevOps
Layer 3: AI & Machine Learning
Layer 2: Backend Engineering & Databases
Layer 1: Programming & Frontend
Get in Touch
/Let's build scalable systems together.
Let's Connect
I am currently open to new opportunities, collaborations, and discussions regarding scalable systems and multilingual AI.