Hello, I'm Suhas
Transforming complex challenges into elegant solutions across AI, software engineering, and cloud infrastructure
About Me
I'm a holistic software engineer.
From low-latency RESTFul APIs to LLM fine-tuning, from backend infrastructure to product development with UI/UX I've shipped across multiple domains and layers.
My core strengths lie in backend engineering, buisiness oriented API engineering, ERP integrations and systems design. BUT I'm not boxed in, I've always worked across the full stack and I'm passionate about post-deployment observability, infrastructure, and scaling.
We are what we repeatedly do. Excellence, then, is not an act, but a habit.
Washington DC cityscape
Experience
Software Engineer
- Designed and integrated RESTful backend APIs supporting analytics, video, and user modules, ensuring data consistency and fault-tolerant communication across mobile platforms
- Automated asynchronous video metadata workflows using AWS Lambda, SQS, and S3, improving scalability and reducing manual processing by 65%
- Configured and maintained CI/CD pipelines with Git, Jenkins, and Terraform, enabling seamless multi-environment deployments and versioned releases to the Play Store
- Collaborated across frontend and backend teams to enhance app performance, implement secure content handling, and achieve 90% feature parity across Android and iOS platforms
Research & Teaching Assistant - AI & Data Science
- Introduced AILA adaptive-attention networks, achieving 30% faster training, 20% lower compute cost, and 15% higher long-range accuracy vs. Transformer/ViT/LSTM baselines
- Explored relevance-aware reward modeling in RLHF; observed gains on datasets with challenges on coding
- Designed ML and GenAI assignments on embeddings, LLM fine-tuning, RAG, feature engineering and EDA
- Reviewed and refined 20+ EDA, Modeling, and explainability techniques (SHAP, LIME) to optimize performance
- Created GenAI tutorials on embeddings, LLM fine-tuning, and RAG, enhancing student learning
- Developed assignments on advanced feature engineering and hyperparameter optimization across ML workflows
Site Reliability Engineer Intern
- Instrumented Golang microservices with Prometheus/Grafana dashboards around Golden Signals (p95/p99 latency, error rate, saturation, availability) + SLO-based alerts. Sustained 99.5% uptime and reduced MTTR by 10%
- Audited AWS Elastic Beanstalk/EC2/RDS and ECS footprint; decommissioned idle DBs/servers and right-sized EC2 based on usage patterns. Optimized capacity with no SLA impact and cut spend by 15%
- Eliminated PostgreSQL RDS version-drift by codifying upgrades in Terraform and Docker: using EOL alerts, blue/green staging, and cutover execution. Achieved compliance, negligible downtime, and removed manual toil
- Modularized Terraform IaC and CloudFormation for VPC, EC2, RDS, IAM, and Beanstalk; integrated plan/apply into CI/CD with policy checks. Standardized environments and cut manual operations by 40%
- Improved application response latency by 25% and scaled throughput by 2× while reducing security vulnerabilities by 30% by tuning critical paths, refactoring bottlenecks, and hardening infrastructure
Software Engineer
- Shipped 20+ RESTful and GraphQL Java Spring Boot/MVC APIs (property listings, realtor profiles, user activity telemetry), complemented by Python serverless jobs on AWS Lambda. Progressively carved domains out of the monolith to improve service ownership and reliability
- Built React/Angular user flows backed by Spring Boot APIs; A/B-tested sign-up/onboarding variants and tightened validation/API contracts. Lifted conversion by 20% and reduced friction across devices
- Led Python ETL for Salesforce CRM (REST/Bulk APIs); mapped schemas, validated data, and loaded to Postgres using idempotent retries. Improved migration speed by 35% while preserving integrity
- Split targeted domains into AWS Lambda microservices behind API Gateway; parameterized configs and wired CI/CD for safe, frequent releases. Achieved 3x deployment frequency and 50% scalability gains
- Implemented user activity-driven lifecycle messaging with AWS SES, Mailchimp, SQS/SNS, and Lambda to notify clients/realtors and trigger follow-ups with event-driven microservices. Reduced customer churn by 15%
- Built Tableau dashboards fed by Jira/Kanban data (defect aging, escape rate, cycle time, WIP, lead time). Accelerated defect triage and enabled QA/Product decisions 30% faster
Projects
Movie Facts App
Full-stack Next.js app with Google OAuth, AI-powered movie trivia, and advanced caching. Features smart rate limiting, security-first architecture, and PostgreSQL database with 10 daily AI calls per user.
Distributed Mini Data Harmonizer
Built a Python-Go distributed pipeline with real concurrency for healthcare data cleaning. Features concurrent processing, REST API, and observability with sub-2-second response times.
AlignAI
Developed and open-sourced the first and only dataset as well as RoBERTa based AI EU Act compliance model, enabling automated detection of PII, bias, and regulation breaches.
QueryMaster AI
AI app to achieve over 90% accuracy in translating natural language into SQL queries, enhancing data accessibility for non-technical users.
SafeHomeSeeker AI
DC SafeHouseFinder AI: Built during aiXplain's Hackathon, provides personalized housing recommendations using AI and geospatial tech.
MatchPoint: The Resume Analyst
Developed a resume scoring app using React and FastAPI, Dockerized deployment on AWS ECS, providing real-time NLP-based feedback and suggestions
Nutrition Tracking Database
Developed a Nutrition Tracking System with SQL Server, Tableau dashboards, and secure handling of user-specific dietary and fitness data.
Research
Waste Management in Urban Localities: An IoT and Machine Learning Solution
Authors: Dr. Viswanatha V, Sreeteja Thummula, Rony Joseph, Varun Raveendra, Suhas K M
Developed an innovative IoT and ML-based solution for waste management in urban localities, improving efficiency and sustainability of municipal waste collection systems.
AILA: Adaptive Integrated Layered Attention
Authors: William Claster, Suhas K M, Dhairya Gundechia
Proposed a novel neural architecture for adaptive integrated layering of attention mechanisms, enabling efficient and effective integration of multiple modalities in AI models.
REALM: Enhancing Reward Models for LLM Alignment introducing Explicit Prompt-Answer Relevance
Authors: William Claster, Suhas K M, Dhairya Gundechia
A research attempt to develop a novel alignment learning model for relevance-enhanced/fine-tuned AI models, enabling improved performance and accuracy in AI applications.
Skills
programming
backend
frontend
cloud
databases
observability
Data Science & AI
practices
Contact me
Fill out the form below and I'll get back to you as soon as possible.
Location
Washington D.C
Open to relocate for on-site opportunities