Abhik Varma Email & Phone Number

Bengaluru, Ka, In

• Architected our core workflow builder for AI agents that leverages node outputs and enables RAG across different data sources.• Built an integration adapter that reduced integration development time from 20 hours to 4 hours, speeding up customer onboarding.• Designed and developed a comprehensive billing system to facilitate self-sign-ups and automated customer billing. Robustly handles 94% of revenue with clear audit logs. A two-phase commit reservation system is used to ensure transactional consistency across distributed services.• Enabled distributed tracing (OpenTelemetry) and metrics (Micrometer) for the event-driven systems, enabling faster identification and resolution of production issues.• Developed an event-driven query-based alerting system capable of scanning 50M rows across databases, APIs and sheets daily to enrich alerts. Apache Drill was used to query non-relational data sources.• Reduced Insights API response time from 8s to 800ms (P99) through query optimization and non-blocking reactive (concurrent) code using Mutiny, enabling swift user experiences on the Insights dashboard.• Implemented scalable two-way sync notification system handling peak of 125K messages/day (99% success rate). Enabled automated incident syncing between tools (Freshdesk) and communication platforms (WhatsApp, Slack), eliminating manual updates.

Bangalore Urban, Karnataka, India

Engineered seamless integration between Sling's content backend and Google Assistant, enabling intuitive content discovery• Developed a Room Database to hold channel mapping data and supply it through a Content Provider.• Developed an interface between Google Assistant and Sling's content backend for improved results from voice commands.• Built an intent logger for debugging Android intents.

Accelerated social affinity data consolidation 25% via multithreading. Built pipelines, monitoring, and analytics for large-scale data ingestion• Through the use of Multithreading and Placekeys, sped up the process of consolidating data from different vendors by ~25%.• Created and maintained Airflow pipelines for daily pre-processing and ingestion of data from S3 buckets to SQL databases. • Built dashboards using Superset to monitor ingested data and notify over Slack if abnormalities were detected.

Bengaluru, Karnataka, India

Benchmarked image captioning performance improvements through Reinforcement Learning and Policy Gradient Optimization (over BLEU, ROUGE-L, & METEOR)

Bachelor Of Technology - Btech, Electrical, Electronics And Communications Engineering

Activities and Societies: Pixels

Bachelor Of Technology - Btech, Minors In Computer Science Engineering